CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Software User Forums > ANSYS > ANSYS Meshing & Geometry

[ICEM] Wing Mesh error

Register Blogs Members List Search Today's Posts Mark Forums Read

Like Tree1Likes
  • 1 Post By shereez234

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   June 15, 2020, 12:45
Default Wing Mesh error
  #1
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
Dear all,

I am a new in ICEM CFD and I am trying to do a mesh for a RANS analysis for the wing attached in the images.

CFDduda.PNG

In order to do so, I have gone through a similar process than this tutorial https://www.youtube.com/watch?v=r5OE9WWmKqI&t=349s. Basically, I am doing an O grid dividing the wing into two blocks, one before the yehudi break and one after it. The mesh appears to be okay in terms of determinant 3x3x3 and quality.

CFDduda2.PNG


However, when I check the mesh I have penetrating elements in some sections specially near LE and TE.

CFDduda3.PNG

The questions are the following:

- This kind of error is a product of bad defining the blocks in pre-meshing step? One solution I thought is that I have to associate the edges along the span with the boundaries of the geometry as in this picture they are not associated:


CFDduda4.PNG
CFDduda5.jpg

However I tried this an the error does not remove at all. It improves but still I have the order of a thousand elements penetrating the surface of the wing.

- This error can be removed with meshing tools such as smooth, merging nodes, etc? If so, what is the solution?




Thank you
Enrique_CFD is offline   Reply With Quote

Old   June 16, 2020, 10:03
Default
  #2
Senior Member
 
M
Join Date: Dec 2017
Posts: 268
Rep Power: 5
AtoHM is on a distinguished road
I think its an issue of association.



A bit hard to tell from the images but I think your block "inside" the wing is quite small and then sometimes ICEM doesn't know where to associate the small surfaces to. It looks like the cells are just oriented along the block face/edge.


If you have done it similarly to the video, you can associate the edges of your block with the geometry edge and ALSO associate that free block face with the wing surfaces. I hope this will fix it.
Attached Images
File Type: jpg assoc.jpg (21.1 KB, 5 views)
AtoHM is offline   Reply With Quote

Old   June 16, 2020, 13:49
Default
  #3
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
Thank you for your comment!!

Just in case, you are refering to the command: Blocking Associations > Associate Face -> Surface?


Association.PNG


I have done that and the problem vanishes in the root. However in the tip there is still a zone in which I am having the problem. Do you have any idea of what is happening there?

Association2.PNG
Enrique_CFD is offline   Reply With Quote

Old   June 17, 2020, 01:26
Default
  #4
Senior Member
 
M
Join Date: Dec 2017
Posts: 268
Rep Power: 5
AtoHM is on a distinguished road
Yes I was refering to that command.

Could you post some closer-up images? Its hard to see anything really. Maybe hide the outer blocks by index control too, so only the relevant lines are visible.
(Index control: make a RMB click onto blocking in the model tree and select index control. There you can hide blocks by index in all 3 spatial directions and o-grids separately. Very handy.)


Also you can check thoroughly again what your edge associations look like (they might be messed up a bit). RMB click on Blocking->Edges and "Show Association".
AtoHM is offline   Reply With Quote

Old   June 17, 2020, 13:38
Default
  #5
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
Thank you for the advice!

I am still having penetrating elements near the tip. I tried the show association command and it seems strange this two arrows pointing downwards (I attach a picture of the whole wing and then this association with zoom)

RANS.PNG

RANS2.PNG

This means that the association is not properly done?
Enrique_CFD is offline   Reply With Quote

Old   June 17, 2020, 13:40
Default
  #6
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
Sorry, I post a wrong image. This is the same with the zoomed in part highlighted:

InkedRANS_LI.jpg
Enrique_CFD is offline   Reply With Quote

Old   June 17, 2020, 15:26
Default
  #7
Senior Member
 
M
Join Date: Dec 2017
Posts: 268
Rep Power: 5
AtoHM is on a distinguished road
Yes this happens especially when you have a very flat block (like yours), then sometimes both sides get associated to the same face, which is obviously wrong. You can try to delete the association of that edge and see if that helps already, I guess it will.
AtoHM is offline   Reply With Quote

Old   June 18, 2020, 05:33
Default
  #8
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
I dissociate this edge and the previous problem was fixed.
The problem now is that I have penetrating elements near this edge that is dissociated.
Is there any way to associate it correctily? (i.e. to the upper surface, reversing the arrow)
Enrique_CFD is offline   Reply With Quote

Old   June 18, 2020, 07:40
Default
  #9
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
I think the problem of the automatic association is due to the highly dihedral of the wing. Here it is a pic of the wing:

Dihedral.PNG

As can be seen at this zoom image at b/3 (where the box starts), both lines are inside the wing and should be associated one at the upper part and other at the lower part.

dihedral2.PNG

However, due to the dihedral the automatic association is a mess...

dihedral3.PNG

Any idea of how to manually ''revert'' the wrong association?
Enrique_CFD is offline   Reply With Quote

Old   June 18, 2020, 08:41
Default
  #10
Senior Member
 
M
Join Date: Dec 2017
Posts: 268
Rep Power: 5
AtoHM is on a distinguished road
Hm I think the edge to face association is automatic (I remember no option to give the face to associate to). However you can try removing (as you did already) and just try to redo associate to face again and see if it helps. Or, you can drag your edge a bit away from the wing and retry the association to see if it catches the right surface then. I am still fumbling in the dark regarding the images (you know what you see there, for me its just lines^^). can you make like a cross-section with the mesh included (cut plane) and close up where the wrong association is? And where your blade block is and looks with respect to the wing profile, maybe at the tip.


BTW: have you deleted the blade inner block?
AtoHM is offline   Reply With Quote

Old   June 18, 2020, 14:53
Default
  #11
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
Thank you again. The undo association and then redo it does not work. Regarding the images, I attach you this one and I hope it clarifies:


InkedClarification1_LI.jpg


As you can see, there are basically two blocks, the first one goes from the root to b/3 and I have no problem with it, and the second one is the one that presents the problem (circled). I have deleted the blocks inside

However, if I move the vertices in the associations that are done in the airfoils and the arrows are reverted in the logical manner. But finally when I do the check te problem does not disappear and I do not know why... Thank you for your patience
Enrique_CFD is offline   Reply With Quote

Old   June 18, 2020, 14:56
Default
  #12
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
Sorry I forgot to post this picture too. This is the zone that I circled in the previous post when I move the vertices and the arrows have the ''logical'' orientation


clarification2.PNG
Enrique_CFD is offline   Reply With Quote

Old   June 19, 2020, 02:48
Default
  #13
Senior Member
 
M
Join Date: Dec 2017
Posts: 268
Rep Power: 5
AtoHM is on a distinguished road
The latest image helps. Ok so its obviously this association that messes things up. Have you tried my suggestion to disassociate the edge, move it "above" the surface (out of the wing volume) and reassociate to surface? If that doesn't work, it looks like there is a geometry line there closer to the leading edge. Can you just try to associate the edge with that geometry line?


BTW I just now realize, these big blocks under and above the wing have veery small angles approaching the boundary of you domain, you will probably get quite bad elements there. If I saw it correctly, the blocking was done another way in the video and there is definitely a better way to make the basic blocking. That might not help with the association issue at the blade but will generally improve the mesh.
AtoHM is offline   Reply With Quote

Old   June 19, 2020, 09:12
Default
  #14
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
Referring to the first paragraph, I will try it this afternoon and send you the result.

Can you further explain the second paragraph? I don't get your point. You mean that my blocks are too big or that I should have the diagonals of the blocks with other orientations? I mean, the picture of the blocks that I passed is just the inner core of the whole blocking (using index control as you said). Are you suggesting to move the verteces in order to have a more orthogonal intersection between the wing and the diagonals?

The main problem of quality arises after the TE since I have to impose very fine elements in this regions to obtain a proper y+ and I obtain a very bad aspect ratio (near the TE very long across the span and as I go further away due to the 'inflation' very long across the orthogonal direction).
Enrique_CFD is offline   Reply With Quote

Old   June 19, 2020, 13:44
Smile
  #15
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
I tried to redo all the associations, making bigger blocks around the wing and associating. Now, this association looks okey but the problem does not vanish and I do not know what else to do.


Penetration.PNG

Summarizing this picture, I have made the blocks bigger and associate in the right side the edges with the edges of the blunt TE. The divisions in the middle I have them associated with the respective airfoils and the edges on the left side I have them associated to the surface since there is only one edge in the geometry (I have also tried to associate them through an edge-curve association but leads to more penetrating elements in this zone). Shouldn't it be enough to properly associate to avoid this penetrating elements? Regardless on how well the curvature is going to be captured by the mesher (I prefer to debug this first and later focus on quality issues)


Regarding what you said about the blocks, I changed its shape a little to achieve a more orthogonal angle. Do you think this is enough?

Penetration2.PNG

Thank you again for being so supportive!
Enrique_CFD is offline   Reply With Quote

Old   June 19, 2020, 17:39
Default
  #16
Senior Member
 
shereez234's Avatar
 
MS
Join Date: Jan 2014
Location: England
Posts: 313
Blog Entries: 1
Rep Power: 9
shereez234 is on a distinguished road
Quote:
Originally Posted by Enrique_CFD View Post
Referring to the first paragraph, I will try it this afternoon and send you the result.

Can you further explain the second paragraph? I don't get your point. You mean that my blocks are too big or that I should have the diagonals of the blocks with other orientations? I mean, the picture of the blocks that I passed is just the inner core of the whole blocking (using index control as you said). Are you suggesting to move the verteces in order to have a more orthogonal intersection between the wing and the diagonals?

The main problem of quality arises after the TE since I have to impose very fine elements in this regions to obtain a proper y+ and I obtain a very bad aspect ratio (near the TE very long across the span and as I go further away due to the 'inflation' very long across the orthogonal direction).

Hello, sorry to jump in atom, i just thought i will share some experience from my side too.

For a good wing structured mesh:

-> Try to keep the blocking simple, do not over complicate it.

-> If you need less elements in trailing edge move the edges that are in the leading edge close to leading edge. Having too many elements in the trailing edge will cause some projection problems. If i remember correctly, a minimum of 8 should do for the trailling edge.

-> Make an extra split close to the tip of the wing so that the round curvature can be captured.

-> Use split edges -> linear for trailing edges and use these edges to link to the edges in the leading edge to capture the dihedral of the wing.

-> High aspect ratio problem is not a problem for boundary layer as long as the values are not too great. A feasible engineering solution cannot be brought forward if aspect ratio is resolved very fine unless you have access to like a good super computer.

-> High aspect ratio is a problem for the far field. You can try to manually set these edges later by copying the edge spacing to selected edges i.e. make them wider along with their length (rather than long thin hexahedrals or quads if 2d)

-> Cell Area and volume transition is quite important (even in the far field).

- > if the cell sizes are very fine use settings->hexa meshing -> projection limit to control the grid quality and projection.

Regards
MS
AtoHM likes this.
shereez234 is offline   Reply With Quote

Old   June 20, 2020, 14:40
Default
  #17
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
Thank you for the tips! I have seen the youtube link and it has helped me a lot. I will take into account this common errors when meshing. However, the error that I was commenting with Atom does not dissappear:

InkedShreeez6_LI.jpg
Shreeez7.PNG

Do you have any idea of what can be happening? Another thing that I do not understand very much is that what is the purpose of creating a part of solid wing? (I am talking about your tutorial). Because I am not able to visualize the mesh with only this part checked in the display pannel (with the WING part desactivated)

Regarding the tip of do not over complicate the blocking, do you consider this divisions along the span are okay? (First picture only showing the geometry and later pictures show the divisions)

Shreeez4.PNG
Shreeez2.PNG
Shreeez3.PNG
Enrique_CFD is offline   Reply With Quote

Old   June 20, 2020, 15:08
Default
  #18
New Member
 
Anonymous
Join Date: May 2020
Posts: 14
Rep Power: 2
Enrique_CFD is on a distinguished road
In addition to the previous part, I have checked what happens with and without the SOLIDWING part activated (i.e. a part of the 4 blocks inside the wing). Without it the mesh looks okey except for the error of the previous post. But when I create this part and I do the pre-mesh, a strange ''hole'' appears in the lower part of the wing

Shreeez8.PNG
Attached Images
File Type: png Shreeez7.PNG (165.5 KB, 4 views)
Enrique_CFD is offline   Reply With Quote

Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
[OpenFOAM.org] Compile OF 2.3 on Mac OS X .... the patch gschaider OpenFOAM Installation 225 August 25, 2015 19:43
Moving mesh Niklas Wikstrom (Wikstrom) OpenFOAM Running, Solving & CFD 122 June 15, 2014 06:20
Compiling dynamicTopoFvMesh for OpenFOAM 2.1.x Saxwax OpenFOAM Installation 25 November 29, 2013 05:34
Ansys Fluent 13.0 UDF compilation problem in Window XP (32 bit) Yogini Fluent UDF and Scheme Programming 7 October 3, 2012 07:24
How to install CGNS under windows xp? lzgwhy Main CFD Forum 1 January 11, 2011 18:44


All times are GMT -4. The time now is 00:20.