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

Airfoil 2D simulation convergence issue

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

Like Tree3Likes
  • 1 Post By frossi
  • 1 Post By arnie333
  • 1 Post By hiep.nguyentrong

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
Old   September 29, 2019, 20:11
Question Airfoil 2D simulation convergence issue
  #1
Member
 
Ferruccio Rossi
Join Date: Jun 2016
Location: Melbourne, FL USA
Posts: 91
Rep Power: 6
frossi is on a distinguished road
Hi all,

I am simulating low speed flow over a 2D airfoil NACA 0012 in FLUENT. I can't get my simulation to converge.

My residuals plot and lift plot continue to diverge even after 10,000 iterations (drag plot looks fine). I thought I have a pretty good mesh so I am not understanding what is the issue.

I was thinking that the divergence might be due to the vortexes toward the trailing edge of the airfoil (see attached pictures) combined with the fact that I choose steady state simulation instead of transient. This said, I am not exactly sure why the vortexes are generated in the first place, since my flow speed is relatively low (29.21 m/s), I expect it to be laminar, and I am using a symmetric airfoil (NACA 0012) at 0 deg angle of attack. All these factors together led me to believe that a steady state run was ideal for my case since no time related phenomena such as vortexes was going to arise.

- My questions are:
1) Are the vortexes shown in my pictures a correct physical phenomena for my simulation setup, or they should not originate in the first place given my setup?
2) Are the vortexes combined with my steady state setup the reason why my simulation is diverging? Or there are other reasons?
2) Should I run a transient simulation instead? If so, can I go with pseudo-transient? (I don't understand the difference between the two).

See below they key information for my mesh and physics set up:

MESH
- Structured mesh (see attached pictures for details).
- 205,476 elements (all quads since it is a structured mesh)

FLUENT PHYSICS
- Pressure based solver
- Steady state simulation (it is NOT pseudo-transient)
- Laminar model (since my airfoil chord is 25 cm the Re is < 500,000
- Air with constant density
- Velocity Inlet BC (29.21 m/s)
- 0 deg angle of attack
- 0.5 flow Courant number
- Walls BC for top and bottom domain surface
- Pressure outlet for the back of the domain
- 10,000 iterations conducted

Any help is greatly appreciated!
Thank you guys
Attached Images
File Type: jpg Residuals.JPG (54.0 KB, 24 views)
File Type: jpg Lift coeff.JPG (64.3 KB, 18 views)
File Type: jpg Drag coeff.JPG (63.6 KB, 16 views)
File Type: jpg Mesh.jpg (124.9 KB, 27 views)
File Type: jpg Mesh detail 1.jpg (195.7 KB, 24 views)
hiep.nguyentrong likes this.
frossi is offline   Reply With Quote

Old   September 29, 2019, 20:14
Default
  #2
Member
 
Ferruccio Rossi
Join Date: Jun 2016
Location: Melbourne, FL USA
Posts: 91
Rep Power: 6
frossi is on a distinguished road
I wasn't able to attach the pictures from the CFD POST in the previous message, so I attached them here. Note the vortexes generated toward the trailing edge of the airfoil.
Attached Images
File Type: jpg Velocity.JPG (81.9 KB, 20 views)
File Type: jpg Velocity detail 1.JPG (89.1 KB, 18 views)
File Type: jpg Velocity detail 2.JPG (95.4 KB, 20 views)
File Type: jpg Velocity detail 3.jpg (196.2 KB, 21 views)
frossi is offline   Reply With Quote

Old   September 30, 2019, 04:17
Default
  #3
New Member
 
Arnie
Join Date: Mar 2017
Posts: 23
Rep Power: 5
arnie333 is on a distinguished road
Hi Frossi,

I cannot see a close-up of your mesh near the airfoil...Did you check element qualities like "Skewness", "Orthogonality" and "Aspect ratio ?

Further, assuming you are using the FLUENT default values for Air Viscosity and Air Density (i.e. sea level ATM conditions), your Re is roughly ~500000. This is on the edge of Turbulent flow, so I would suggest using a Turbulence Solver (perhaps k-w SST) and leave "Pseudo Transient" enabled.

PS. You can also make the top and bottom wall as Slip Wall (No Shear), although this is more for aesthetics.
frossi likes this.
arnie333 is offline   Reply With Quote

Old   October 4, 2019, 00:43
Default
  #4
New Member
 
Nguyen Trong Hiep
Join Date: Aug 2018
Posts: 12
Rep Power: 4
hiep.nguyentrong is on a distinguished road
can you set min-max distane of your Cd plot close to your magnitude then post here
frossi likes this.
hiep.nguyentrong is offline   Reply With Quote

Old   October 24, 2019, 00:04
Default
  #5
Member
 
Ferruccio Rossi
Join Date: Jun 2016
Location: Melbourne, FL USA
Posts: 91
Rep Power: 6
frossi is on a distinguished road
Arnie,

Thank you very much for your suggestion.
Sorry for the late reply. I conducted a lot of tests during this time.

I tried the k-omega SST model as you suggested and I am getting a solution converged to 1e-09, which I think it's excellent! Do you have some literature that you would recommend reading regarding turbulence models? I would like to be able to justify my choice for this turbulence model.

See below the detailed pictures of the mesh as you asked. I think the mesh looks good. What are your thoughts? Feel free to say if you think improvements are needed.


Now that I was able to solve this first issue, I have these questions:

1) I have a curiosity regarding the residuals plots. For CONTINUITY (black line) and some other residuals, you can see a slight oscillating behavior before the plot flat-lined (see pics). I can't understand what that might mean. In this specific case, I used a mesh with a quite large domain size. With this large domain size, the oscillatory behavior is small, and it dissipates when it flat-lines. When I use a mesh with a smaller domain size, the oscillation is very large, and the residual never flat-lines, as it keeps oscillating, and sometimes it also ends up diverging.

2) I am analyzing the Y plus for this simulation, to determine if it is in the correct range or not. However, I am not sure how to determine this. For my simulation, the Y plus is between 5 and 20 for most of the chord length (see chart). I was reading a lot about this, but people have very different opinions, some saying it needs to be between 1 and 5, some say above 5, some say below 200, etc... . From what I understand (correct me if I am wrong) the Y plus choice varies depending on the turbulence model I choose. I am not sure what would be a way to define if the Y plus range I obtain is good for my simulation or not.

3) For what concerns your last comment you made about making the wall no shear, how would you do it? I don't see an option for that. The only thing I can think of is selecting "specified shear" and inputting 0 for both X shear and Y shear. Is this what you mean?



Thank you for your support.
Attached Images
File Type: jpg mesh 1.jpg (107.9 KB, 9 views)
File Type: jpg mesh 2.jpg (110.2 KB, 6 views)
File Type: jpg mesh 3.jpg (129.9 KB, 8 views)
File Type: png Y plus (1).PNG (27.4 KB, 9 views)
File Type: png Oscillation (1).PNG (49.6 KB, 8 views)
frossi is offline   Reply With Quote

Old   October 24, 2019, 00:13
Default
  #6
Member
 
Ferruccio Rossi
Join Date: Jun 2016
Location: Melbourne, FL USA
Posts: 91
Rep Power: 6
frossi is on a distinguished road
Nguyen,

Thank you for your reply. As you can read above, I was able to get the solution to converge, so my Cd plot now is fully converged (I don't have a picture to upload here at the moment, I would need to re-run the simulation to take a new screenshot).

However, while I obtain a drag force value of 1.75 N (which I believe to be reasonable for my set up) I get a Cd of 2.3, which seems very high. if i calculate Cd from the lift formula, I get 0.012 for Cd, which is way lower than 2.3. Do you know why Fluent is outputting such a high value? I suspect it has to do with the reference values (see pic)? or some other information I input during set up?

Thank you!
Attached Images
File Type: png reference.PNG (57.9 KB, 15 views)
frossi is offline   Reply With Quote

Old   October 24, 2019, 05:40
Default
  #7
New Member
 
Nguyen Trong Hiep
Join Date: Aug 2018
Posts: 12
Rep Power: 4
hiep.nguyentrong is on a distinguished road
Quote:
Originally Posted by frossi View Post
Nguyen,

Thank you for your reply. As you can read above, I was able to get the solution to converge, so my Cd plot now is fully converged (I don't have a picture to upload here at the moment, I would need to re-run the simulation to take a new screenshot).

However, while I obtain a drag force value of 1.75 N (which I believe to be reasonable for my set up) I get a Cd of 2.3, which seems very high. if i calculate Cd from the lift formula, I get 0.012 for Cd, which is way lower than 2.3. Do you know why Fluent is outputting such a high value? I suspect it has to do with the reference values (see pic)? or some other information I input during set up?

Thank you!
u must set ref value to calculate cd (velocity, area, length, rho) or report drag and calculate it
hiep.nguyentrong is offline   Reply With Quote

Old   October 24, 2019, 18:29
Default
  #8
New Member
 
Diogo
Join Date: Apr 2019
Posts: 4
Rep Power: 3
diogoncsa is on a distinguished road
Quote:
Originally Posted by hiep.nguyentrong View Post
u must set ref value to calculate cd (velocity, area, length, rho) or report drag and calculate it
What are the ref values for this case? I am having the same low cd and cl issue with airfoil 2D simulations
diogoncsa is offline   Reply With Quote

Old   October 24, 2019, 21:42
Default
  #9
New Member
 
Nguyen Trong Hiep
Join Date: Aug 2018
Posts: 12
Rep Power: 4
hiep.nguyentrong is on a distinguished road
velocity: 29.21 m/s, length : 0.25, area: 0.25
hiep.nguyentrong is offline   Reply With Quote

Old   October 25, 2019, 10:12
Default
  #10
New Member
 
Diogo
Join Date: Apr 2019
Posts: 4
Rep Power: 3
diogoncsa is on a distinguished road
Quote:
Originally Posted by hiep.nguyentrong View Post
velocity: 29.21 m/s, length : 0.25, area: 0.25
That solved my problem, thanks Nguyen!
diogoncsa is offline   Reply With Quote

Reply

Tags
airfoil, divergence, fluent, residuals, vortex

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
Question about heat transfer simulation convergence Anna Tian CFX 26 November 14, 2017 04:37
A convergence issue of RAE2822 airfoil leejearl SU2 0 May 11, 2016 09:27
Convergence issue in natural convection problem chrisf90 FLUENT 5 March 5, 2016 08:30
Convergence of jet flow simulation MiraLisa FLUENT 0 August 15, 2013 04:44
2D SST Simulation Airfoil - Convergence Problem Kraemer CFX 10 April 16, 2011 07:22


All times are GMT -4. The time now is 21:29.