[ICEM] meshing of a journal bearing in ICEM

Pac · August 23, 2011, 07:20

Dear community,

I am trying to mesh a journal bearing which more or less consists of two pipe surfaces with different diameters (61.32mm and 61.205mm) as you see, the resulting gap is only 0.0575mm high. The geometry was developed in DM.

From the many problems I have already encountered, there are two, which I have no idea where they come from and much less how to solve them.

1. When generating a PreMesh, at some points periodic around the geometry, Icem melds the surfaces. That of course divides the gap and leads to an unusable Mesh.

2. The Mesh has a large amount of inverted volumes and smoothing only leads to more inverted volumes.

I used an O-Grid.

If you need more details, please ask me for them.

I would be very glad if somebody told where the problems result from and how to solve them.
Thank you,
Pac

BrolY · August 24, 2011, 08:32

Did you associate all the edges to the curves?

Post some pictures of your problem, it would help a lot !

Pac · August 24, 2011, 09:03

Thank you for your reply!

yes, I associated all edges to the curves, I also tried it with associating the vertices to points I created along the curves, but that would not help either.

Here are two pictures, one of the geometry and one of the defective mesh, the location where the inner and the outer surface merge, occures periodicly around the cylinder, left and right of the oil groove.

bear1.jpg bear2.jpg

BrolY · August 24, 2011, 09:08

Can you attach your CAO here ?
I can have a look if you want.

Pac · August 24, 2011, 12:40

That would be great. But I must admit that I do not know what you mean with "CAO". Which files do you need?

BrolY · August 24, 2011, 12:43

The *.tin file ! CAO = geometry

Pac · August 24, 2011, 12:54

oh, ok

here it is. Project1.1.zip

BrolY · August 25, 2011, 04:57

Ok I think the issue is the gap between the part "Shaft" and the part "Bear" is very thin. At the same time, your geoemtry is pretty big.
It looks like ICEM has troubles to handle this case. You could try to create a lot of splits and to associate each vertices to a node (and also each curves to each edges) ... It would require a lot of time but some ICEM tools could help to handle that.

Have fun

Pac · August 25, 2011, 07:14

Thank you very much for your help, I already suspected that the dimensions might be a problem but I hoped that there was a trivial solution to this... I will try what you recommended and create more splits and more associations.

Greetings, pac

stuart23 · August 25, 2011, 21:59

You could try decreasing all your tolerences. The tolerences might be low enough that ICEM just believes that you wanted one surface instead of two seperate surfaces in some areas.

Also maybe try spliting surface by curvature to generate more lines for ICEM to read off.

Stu

BrolY · August 26, 2011, 04:19

Stuart, if ICEM believes there is only one surface, you would end with a 2D mesh which is not the purpose of this study I guess ...

Pac · August 27, 2011, 07:57

First of all, thank you two for your help.

After doing what BrolY recommended, ICEM created me an acceptable mesh. There are still some bad elements, but with a little fine tuning, I think I will get rid of them.

I think what Stuart ment is to decrease the tolerence so that ICEM understands that there are two surfaces instead of (in some regions) one. But since it worked by splitting the surfaces, I think that the tolerence was not the problem.

I now ponder if a struct mesh created with the workbench mesher by sweeping would be a better solution than the unstruct mesh created with ICEM...

qingyue1987 · December 12, 2011, 07:56

hi,
i think you should try to us ICEM to creat much more block .then doing the association work .

August 23, 2011, 07:20	meshing of a journal bearing in ICEM	#1
Pac New Member Join Date: Jul 2011 Location: Karlsruhe Posts: 7 Rep Power: 15	Dear community, I am trying to mesh a journal bearing which more or less consists of two pipe surfaces with different diameters (61.32mm and 61.205mm) as you see, the resulting gap is only 0.0575mm high. The geometry was developed in DM. From the many problems I have already encountered, there are two, which I have no idea where they come from and much less how to solve them. 1. When generating a PreMesh, at some points periodic around the geometry, Icem melds the surfaces. That of course divides the gap and leads to an unusable Mesh. 2. The Mesh has a large amount of inverted volumes and smoothing only leads to more inverted volumes. I used an O-Grid. If you need more details, please ask me for them. I would be very glad if somebody told where the problems result from and how to solve them. Thank you, Pac

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[ICEM] complex 2D meshing on ICEM	kassab	ANSYS Meshing & Geometry	26	April 14, 2014 22:02
simulation of a journal bearing in Comsol	Pac	COMSOL	0	July 6, 2011 11:59
[ICEM] ICEM meshing problem	xyq102296	ANSYS Meshing & Geometry	6	October 28, 2010 11:09
ICEM Meshing	Jaloha	CFX	0	April 1, 2008 07:24
Missing tets along line when meshing with ICEM CFD	Georges P. Côté	CFX	6	March 23, 2006 01:34

August 24, 2011, 08:32		#2
BrolY Senior Member AB Join Date: Sep 2009 Location: France Posts: 323 Rep Power: 22	Did you associate all the edges to the curves? Post some pictures of your problem, it would help a lot !

August 24, 2011, 09:03		#3
Pac New Member Join Date: Jul 2011 Location: Karlsruhe Posts: 7 Rep Power: 15	Thank you for your reply! yes, I associated all edges to the curves, I also tried it with associating the vertices to points I created along the curves, but that would not help either. Here are two pictures, one of the geometry and one of the defective mesh, the location where the inner and the outer surface merge, occures periodicly around the cylinder, left and right of the oil groove. bear1.jpg bear2.jpg

August 24, 2011, 09:08		#4
BrolY Senior Member AB Join Date: Sep 2009 Location: France Posts: 323 Rep Power: 22	Can you attach your CAO here ? I can have a look if you want.

August 24, 2011, 12:40		#5
Pac New Member Join Date: Jul 2011 Location: Karlsruhe Posts: 7 Rep Power: 15	That would be great. But I must admit that I do not know what you mean with "CAO". Which files do you need?

August 24, 2011, 12:43		#6
BrolY Senior Member AB Join Date: Sep 2009 Location: France Posts: 323 Rep Power: 22	The *.tin file ! CAO = geometry

August 24, 2011, 12:54		#7
Pac New Member Join Date: Jul 2011 Location: Karlsruhe Posts: 7 Rep Power: 15	oh, ok here it is. Project1.1.zip

August 25, 2011, 04:57		#8
BrolY Senior Member AB Join Date: Sep 2009 Location: France Posts: 323 Rep Power: 22	Ok I think the issue is the gap between the part "Shaft" and the part "Bear" is very thin. At the same time, your geoemtry is pretty big. It looks like ICEM has troubles to handle this case. You could try to create a lot of splits and to associate each vertices to a node (and also each curves to each edges) ... It would require a lot of time but some ICEM tools could help to handle that. Have fun

August 25, 2011, 07:14		#9
Pac New Member Join Date: Jul 2011 Location: Karlsruhe Posts: 7 Rep Power: 15	Thank you very much for your help, I already suspected that the dimensions might be a problem but I hoped that there was a trivial solution to this... I will try what you recommended and create more splits and more associations. Greetings, pac

August 25, 2011, 21:59		#10
stuart23 Senior Member Stuart Buckingham Join Date: May 2010 Location: United Kingdom Posts: 267 Rep Power: 26	You could try decreasing all your tolerences. The tolerences might be low enough that ICEM just believes that you wanted one surface instead of two seperate surfaces in some areas. Also maybe try spliting surface by curvature to generate more lines for ICEM to read off. Stu

August 26, 2011, 04:19		#11
BrolY Senior Member AB Join Date: Sep 2009 Location: France Posts: 323 Rep Power: 22	Stuart, if ICEM believes there is only one surface, you would end with a 2D mesh which is not the purpose of this study I guess ...

August 27, 2011, 07:57		#12
Pac New Member Join Date: Jul 2011 Location: Karlsruhe Posts: 7 Rep Power: 15	First of all, thank you two for your help. After doing what BrolY recommended, ICEM created me an acceptable mesh. There are still some bad elements, but with a little fine tuning, I think I will get rid of them. I think what Stuart ment is to decrease the tolerence so that ICEM understands that there are two surfaces instead of (in some regions) one. But since it worked by splitting the surfaces, I think that the tolerence was not the problem. I now ponder if a struct mesh created with the workbench mesher by sweeping would be a better solution than the unstruct mesh created with ICEM...

December 12, 2011, 07:56		#13
qingyue1987 New Member JIM Join Date: Nov 2011 Posts: 5 Rep Power: 15	hi, i think you should try to us ICEM to creat much more block .then doing the association work .