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

[ICEM] How to create prism layers on sharp trailing edges in ICEM CFD (3D Tetra Mesh)?

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

Like Tree4Likes
  • 2 Post By kad
  • 1 Post By kad
  • 1 Post By kad

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   October 22, 2018, 11:27
Default How to create prism layers on sharp trailing edges in ICEM CFD (3D Tetra Mesh)?
  #1
Member
 
MWRS
Join Date: Apr 2018
Posts: 95
Rep Power: 3
waseeqsiddiqui is on a distinguished road
Regarding prism meshing around sharp trailing edges in ICEM, I have searched everywhere, tried every possible setting combination. Every related discussion in CFD forums ends up into hexa meshing. Please could someone help me with this problem. The quality and angles/ aspect ratio of volume tetras from the last prism layer in TEs are pretty bad almost everytime. Please help.
waseeqsiddiqui is offline   Reply With Quote

Old   October 22, 2018, 16:57
Default
  #2
Senior Member
 
Gert-Jan
Join Date: Oct 2012
Posts: 899
Rep Power: 15
Gert-Jan is on a distinguished road
- which method did you use in ICEM? Octree? Fluent Meshing? Delauney?
- is your initial surface mesh ok everywhere around your trailling edge?
- did you read the help in Global Mesh Setup/Prism Meshing Parameters/Global Prisms Settings? Did you use Max Prism Angle = 180 and Prism Height Limit Factor = 0.5
- which CFD-solver will you use eventually?
Gert-Jan is offline   Reply With Quote

Old   October 22, 2018, 23:07
Default
  #3
Member
 
MWRS
Join Date: Apr 2018
Posts: 95
Rep Power: 3
waseeqsiddiqui is on a distinguished road
Thankyou for responding. Octree.
Yes. I had defined thin cuts so it captured the surface completely and contains very good quality surface tris.

Yes i have been reading the prism help manual and I have tried 180 and 0.5. I have tried many other combos like allowing only pyramids at sharp edges by setting 160 and max height over base. Still no luck.
Ansys Fluent.
waseeqsiddiqui is offline   Reply With Quote

Old   October 23, 2018, 03:47
Default
  #4
Senior Member
 
Gert-Jan
Join Date: Oct 2012
Posts: 899
Rep Power: 15
Gert-Jan is on a distinguished road
- How does you trailing edge look like? Can you decrease the element size?
- Who says your mesh is not ok? Your supervisor? Fluent? ICEM?

- Did you you try the Pre-inflation (fluent) method?- Did you try to smooth the mesh in Fluent?
- Fluent and Tetra-mesh do not provide the best solutions. Maybe convert it to Polyhedra mesh?
- Maybe try CFX? CFX is much nore foregiven than Fluent when it comes to bad elements............
Gert-Jan is offline   Reply With Quote

Old   October 23, 2018, 08:16
Default
  #5
Member
 
MWRS
Join Date: Apr 2018
Posts: 95
Rep Power: 3
waseeqsiddiqui is on a distinguished road
Mesh diverges in Fluent.
Only the first tetra elements after the last prism layer at the sharp bend of TE are having good skewness but poor quality and angles.
I tried edges and surface meshing. No luck.
waseeqsiddiqui is offline   Reply With Quote

Old   October 23, 2018, 08:19
Default
  #6
Member
 
MWRS
Join Date: Apr 2018
Posts: 95
Rep Power: 3
waseeqsiddiqui is on a distinguished road
Nope didn't Fluent Pre-inflation method? I tried smoothing the mesh in fluent though.
Not supervisors, its Ansys Fluent.
Will try CFX
waseeqsiddiqui is offline   Reply With Quote

Old   October 23, 2018, 08:21
Default
  #7
Senior Member
 
Gert-Jan
Join Date: Oct 2012
Posts: 899
Rep Power: 15
Gert-Jan is on a distinguished road
CFX is a lifesaver in this respect. Unless you have negative elements.
Gert-Jan is offline   Reply With Quote

Old   October 23, 2018, 08:59
Default
  #8
kad
Senior Member
 
Join Date: Feb 2013
Location: Germany
Posts: 200
Rep Power: 19
kad will become famous soon enoughkad will become famous soon enough
In case of tetra/prism meshing for e.g. naca airfoil you can use something like auxiliary surface for prism generation. I made a little drawing of that idea. The auxiliary surface (yellow) leads the prism layer all the way out of the domain and avoids critical angles in the mesh. The auxiliary surface has to be defined as interior in your solver then.
Attached Images
File Type: jpg nacaPrism1.jpg (26.2 KB, 66 views)
lgn_aero and waseeqsiddiqui like this.
kad is offline   Reply With Quote

Old   October 26, 2018, 12:18
Default
  #9
Member
 
MWRS
Join Date: Apr 2018
Posts: 95
Rep Power: 3
waseeqsiddiqui is on a distinguished road
Quote:
Originally Posted by kad View Post
In case of tetra/prism meshing for e.g. naca airfoil you can use something like auxiliary surface for prism generation. I made a little drawing of that idea. The auxiliary surface (yellow) leads the prism layer all the way out of the domain and avoids critical angles in the mesh. The auxiliary surface has to be defined as interior in your solver then.
Good idea! So my guess is that ICEM calls this auxilary surface as internal wall! This won't cause any error in results right?
waseeqsiddiqui is offline   Reply With Quote

Old   November 2, 2018, 13:49
Default
  #10
Far
Super Moderator
 
Sijal
Join Date: Mar 2009
Location: Islamabad
Posts: 4,538
Blog Entries: 6
Rep Power: 49
Far has a spectacular aura aboutFar has a spectacular aura about
Send a message via Skype™ to Far
Quote:
Originally Posted by kad View Post
In case of tetra/prism meshing for e.g. naca airfoil you can use something like auxiliary surface for prism generation. I made a little drawing of that idea. The auxiliary surface (yellow) leads the prism layer all the way out of the domain and avoids critical angles in the mesh. The auxiliary surface has to be defined as interior in your solver then.
What happens at the end of auxiliary surface?
Far is offline   Reply With Quote

Old   November 9, 2018, 10:24
Default
  #11
kad
Senior Member
 
Join Date: Feb 2013
Location: Germany
Posts: 200
Rep Power: 19
kad will become famous soon enoughkad will become famous soon enough
It extends through the whole rest of the domain so it spans a "nice" 90 degree angle at e.g. outlet. Okay, it is not an very elegant solution but it is one possible way to avoid bad prism elements.
waseeqsiddiqui likes this.
kad is offline   Reply With Quote

Old   November 9, 2018, 10:27
Default
  #12
kad
Senior Member
 
Join Date: Feb 2013
Location: Germany
Posts: 200
Rep Power: 19
kad will become famous soon enoughkad will become famous soon enough
Quote:
Originally Posted by waseeqsiddiqui View Post
Good idea! So my guess is that ICEM calls this auxilary surface as internal wall! This won't cause any error in results right?

Usually the elements must be defined as internal in the solver, too. This won't give you any error because these faces are then treated like any other internal face.
waseeqsiddiqui likes this.
kad is offline   Reply With Quote

Old   November 9, 2018, 11:27
Default
  #13
Member
 
MWRS
Join Date: Apr 2018
Posts: 95
Rep Power: 3
waseeqsiddiqui is on a distinguished road
Interesting
waseeqsiddiqui is offline   Reply With Quote

Old   January 6, 2019, 15:38
Default
  #14
New Member
 
N/A
Join Date: Mar 2013
Posts: 16
Rep Power: 8
lgn_aero is on a distinguished road
Quote:
Originally Posted by kad View Post
Usually the elements must be defined as internal in the solver, too. This won't give you any error because these faces are then treated like any other internal face.
Otherwise, after meshing (prism) completion change the association (remove) so that this internal sarface will be no more so no additional setting in solver tooooo.

Last edited by lgn_aero; January 8, 2019 at 10:43.
lgn_aero is offline   Reply With Quote

Old   January 30, 2019, 11:56
Default Attempted with internal wall: yet getting worse
  #15
Member
 
MWRS
Join Date: Apr 2018
Posts: 95
Rep Power: 3
waseeqsiddiqui is on a distinguished road
Quote:
Originally Posted by kad View Post
In case of tetra/prism meshing for e.g. naca airfoil you can use something like auxiliary surface for prism generation. I made a little drawing of that idea. The auxiliary surface (yellow) leads the prism layer all the way out of the domain and avoids critical angles in the mesh. The auxiliary surface has to be defined as interior in your solver then.
Dear Kad,
I am uploading the snaps for my experiment of thin baffle on a wing with sharp trailing edge. The surfaces of upper and lower wing are named WING UPPER and WING LOWER. Thin Cuts are defined to ensure the surfaces are captured accurately at edge criterion 0.2. Points of Trailing Edge are copied and translated out of the domain. Single baffle surface is created by connecting these four points and setting the part as internal wall using mesh by parts setup.(Figure 3). Prism settings are kept as default; stair stepping enabled, 180 max prism angle, min prism quality was set once to 0.01 and once as 1e-06. Changes are made in fillet ratio.


Observation: Quality elements are even worse with the internal surface, mainly on each of the edges of the thin surface just like the way they get worse at a sharp trailing edge (high aspect ratio). All my attempts of changing prism settings shows the same behaviour. Even Stair Stepping is enabled and disabled as stated by Mr. Simon. Furthermore, the prisms, once they end at sharp TE and begin at internal surface; they converges into an undesirable height.

What am I doing wrong? Help needed.
Thanks.
Attached Images
File Type: jpg 1.jpg (91.6 KB, 22 views)
File Type: jpg 2.jpg (187.4 KB, 25 views)
File Type: jpg 3.jpg (169.4 KB, 28 views)
File Type: jpg 4.jpg (200.6 KB, 25 views)
File Type: jpg 5.jpg (112.9 KB, 18 views)

Last edited by waseeqsiddiqui; January 30, 2019 at 12:00. Reason: title editing
waseeqsiddiqui is offline   Reply With Quote

Old   January 30, 2019, 11:59
Default Additional Snaps
  #16
Member
 
MWRS
Join Date: Apr 2018
Posts: 95
Rep Power: 3
waseeqsiddiqui is on a distinguished road
Quote:
Originally Posted by waseeqsiddiqui View Post
Dear Kad,
I am uploading the snaps for my experiment of thin baffle on a wing with sharp trailing edge. The surfaces of upper and lower wing are named WING UPPER and WING LOWER. Thin Cuts are defined to ensure the surfaces are captured accurately at edge criterion 0.2. Points of Trailing Edge are copied and translated out of the domain. Single baffle surface is created by connecting these four points and setting the part as internal wall using mesh by parts setup.(Figure 5). Prism settings are kept as default; stair stepping enabled, 180 max prism angle, min prism quality was set once to 0.01 and once as 1e-06. Changes are made in fillet ratio.


Observation: Quality elements are even worse with the internal surface, mainly on each of the edges of the thin surface just like the way they get worse at a sharp trailing edge (high aspect ratio). All my attempts of changing prism settings shows the same behaviour. Even Stair Stepping is enabled and disabled as stated by Mr. Simon. Furthermore, the prisms, once they end at sharp TE and begin at internal surface; they converges into an undesirable height.

What am I doing wrong? Help needed.
Thanks.

Few more snaps added
Attached Images
File Type: jpg 6.jpg (197.1 KB, 27 views)
File Type: jpg 7.jpg (94.7 KB, 24 views)
waseeqsiddiqui 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
How to mesh sharp corners in ICEM CFD abhilasht ANSYS Meshing & Geometry 1 January 8, 2019 10:53
how to deal with sharp trailing edges in icem joegi.geo ANSYS Meshing & Geometry 31 March 5, 2018 20:57
Inner geometry gets lost exporting mesh from ICEM CFD to CFX-Pre powpow CFX 3 December 20, 2012 10:14
[ANSYS Meshing] ICEM CFD create mesh for OpenFoam cyclic danilo ANSYS Meshing & Geometry 0 August 14, 2012 06:58
ICEM clustering of mesh points and multiple edges webhaw CFX 4 July 25, 2007 17:29


All times are GMT -4. The time now is 00:22.