[ICEM] Strategy for sharp corners

RPFigueiredo · March 22, 2013, 09:00

Hello again,

I've managed to create a relatively satisfactory mesh, but I am struggling with its quality. As you can see in the image bellow, the model contains to problematic corners - very obtuse/acute angles.

You any of you recommend a different way of meshing this in order to improve the quality? Above 0.4 would be OK!

Thank you in advance!

Far · March 22, 2013, 09:30

I am not able to see any image ... please attach again

RPFigueiredo · March 22, 2013, 10:00

Sorry, here it is!

Far · March 22, 2013, 10:03

what is the problem?

RPFigueiredo · March 22, 2013, 10:10

The element quality near the outlet and the periodic walls is very low due to the angles, I guess. Do you reckon there's a better way to mesh those corners, or is the low quality something I'll have to accept?

Far · March 22, 2013, 10:16

Actually I cannot visualize that area. Can you pin point that area and how much quality do you get?

RPFigueiredo · March 22, 2013, 10:27

Here. Basically the geometry itself causes low orthogonality.

Far · March 22, 2013, 10:30

How much low? Like how much quality you have, min angle, dihedral angle, orthogonality, max volume change?

Plot with quality check menu and click on bars with low quality and post those pics

RPFigueiredo · March 22, 2013, 10:47

Since I've just started dealing with CFD, I have been following the >0.4 Quality index (it's a CFX parameter). Not really sure how much would the max. volume change figure be, though!
I've also been informed to keep angles >18deg.

Thanks a lot for your time!

BrolY · March 22, 2013, 10:58

Quality > 0.3 is good enough.
But maybe you could improve the quality in this area. Can you share .tin and .blk files ?

Far · March 22, 2013, 11:10

I can see no problem in mesh. Every things seems ok. What is the maximum Y+?

RPFigueiredo · March 22, 2013, 11:16

Not sure, I'll carry out a sensibility analysis probably resulting in a finner mesh (~1.5m elements, perhaps).

Thanks for the replies!

RPFigueiredo · March 22, 2013, 11:17

Quote:

Originally Posted by BrolY

Quality > 0.3 is good enough.
But maybe you could improve the quality in this area. Can you share .tin and .blk files ?

I've tried doing it by moving the nodes - no success!
Is there another method you'd suggest? I was advised not to share the files, sorry! --'

Far · March 22, 2013, 11:25

Quote:

Originally Posted by RPFigueiredo

I was advised not to share the files, sorry! --'

In fact sharing them will give you good response from forum fellows. You can share privately.

Choice is yours.

RPFigueiredo · March 22, 2013, 11:45

OK, If I would get some help with that, I could share the files for both stator and rotor (which looks a bit more messy...).
Can I share them here or should I get your e-mails through PM?

Far · March 22, 2013, 11:46

if it is confidential then through pm.

March 22, 2013, 09:00	Strategy for sharp corners	#1
RPFigueiredo Member Rolando Figueiredo Join Date: Mar 2013 Posts: 41 Rep Power: 13	Hello again, I've managed to create a relatively satisfactory mesh, but I am struggling with its quality. As you can see in the image bellow, the model contains to problematic corners - very obtuse/acute angles. You any of you recommend a different way of meshing this in order to improve the quality? Above 0.4 would be OK! Thank you in advance!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
[ICEM] Mesh strategy large tank	DoubleG	ANSYS Meshing & Geometry	2	January 10, 2013 11:15
[ICEM] Blocking Strategy for Intersecting Tapered Cylinders	scribby182	ANSYS Meshing & Geometry	2	December 17, 2012 11:02
How to find a good time scale strategy?	StefanG	CFX	19	June 8, 2012 09:41
Strategy for keeping track of files and folders?	Dre	CFX	1	October 30, 2008 14:00
Poinsot and Lele Characteristic BCS strategy for NS equations	Ravikanth Avancha	Main CFD Forum	1	March 15, 1999 20:51

March 22, 2013, 09:30		#2
Far Senior Member Sijal Join Date: Mar 2009 Location: Islamabad Posts: 4,558 Blog Entries: 6 Rep Power: 54	I am not able to see any image ... please attach again

March 22, 2013, 10:03		#4
Far Senior Member Sijal Join Date: Mar 2009 Location: Islamabad Posts: 4,558 Blog Entries: 6 Rep Power: 54	what is the problem?

March 22, 2013, 10:10		#5
RPFigueiredo Member Rolando Figueiredo Join Date: Mar 2013 Posts: 41 Rep Power: 13	The element quality near the outlet and the periodic walls is very low due to the angles, I guess. Do you reckon there's a better way to mesh those corners, or is the low quality something I'll have to accept?

March 22, 2013, 10:16		#6
Far Senior Member Sijal Join Date: Mar 2009 Location: Islamabad Posts: 4,558 Blog Entries: 6 Rep Power: 54	Actually I cannot visualize that area. Can you pin point that area and how much quality do you get?

March 22, 2013, 10:30		#8
Far Senior Member Sijal Join Date: Mar 2009 Location: Islamabad Posts: 4,558 Blog Entries: 6 Rep Power: 54	How much low? Like how much quality you have, min angle, dihedral angle, orthogonality, max volume change? Plot with quality check menu and click on bars with low quality and post those pics

March 22, 2013, 10:58		#10
BrolY Senior Member AB Join Date: Sep 2009 Location: France Posts: 323 Rep Power: 22	Quality > 0.3 is good enough. But maybe you could improve the quality in this area. Can you share .tin and .blk files ?

March 22, 2013, 11:10		#11
Far Senior Member Sijal Join Date: Mar 2009 Location: Islamabad Posts: 4,558 Blog Entries: 6 Rep Power: 54	I can see no problem in mesh. Every things seems ok. What is the maximum Y+?

March 22, 2013, 11:16		#12
RPFigueiredo Member Rolando Figueiredo Join Date: Mar 2013 Posts: 41 Rep Power: 13	Not sure, I'll carry out a sensibility analysis probably resulting in a finner mesh (~1.5m elements, perhaps). Thanks for the replies!

March 22, 2013, 11:45		#15
RPFigueiredo Member Rolando Figueiredo Join Date: Mar 2013 Posts: 41 Rep Power: 13	OK, If I would get some help with that, I could share the files for both stator and rotor (which looks a bit more messy...). Can I share them here or should I get your e-mails through PM?

March 22, 2013, 11:46		#16
Far Senior Member Sijal Join Date: Mar 2009 Location: Islamabad Posts: 4,558 Blog Entries: 6 Rep Power: 54	if it is confidential then through pm.