CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Steady level flight airfoil analysis using ANSYS CFX v12 (http://www.cfd-online.com/Forums/cfx/73734-steady-level-flight-airfoil-analysis-using-ansys-cfx-v12.html)

 flutter March 15, 2010 17:12

Steady level flight airfoil analysis using ANSYS CFX v12

Hello,

I have been working on doing a steady state FSI analysis using CFX v12 of an airfoil. The mesh is very coarse, and I am having troubles with convergence.

I was wondering if someone could please point me in the right direction as to how I should be setting this problem up.

Here is a list of the things I have already tried:
1. Analysis type: Ansys Multifield, Boundary Conditions: Inlet, wall, symmetry, time step:1s, total time: 10s

2. Analysis type: Multifield, Boundary Conditions: Opening, wall, symmetry

3. Defining separate domains for the fluid as well as the solid and then same as above.

Any help would be greatly appreciated.

 toil March 15, 2010 22:09

I haven't do any FSI simulation before, but according to my transient computing experience, your time step maybe too long for FSI problem, I do have a mate who was working on FSI simulation, a rectangular wing precisely, his time step is equal to 1E-3 magnitude.

 Shakeel April 6, 2010 06:27

Dear flutter,

I am working on a very similar project, I am also trying to simulate the aeroelastic behaviour of an aerofoil using CFX v12. I believe my structural model is not valid. (Hence I receive an error during the solving stage i.e. a read error between the data exchange) I am trying to run a 2D model with a unit span and an open section on the non-supported side. I used shell elements to represent the skin of the wing(therefore you could say inside the aerofoil is a region of emptiness). I have successfully run the FSI problem using a solid aerofoil (for testing the principles) but this is obviously not physically valid. Could I please ask how you modelled the aerofoil in the mechanical section? Was it a 3D aerofoil? Lastly, I am importing the model using an IGES file but if there is a fully closed section the geometry appears as a solid and I cannot work out how to make it a shell structure, did you have a similar problem? The CFD section has been validated and I am confident the steady dynamics are working fine.

Please criticise as much as possible as I am very new to FSI and need all the help I can get. If you could maybe direct me to some tutorials so I can help myself that would also be appreciated.

Thank you in advance and I hope to hear from you soon.

Shakeel.

This is the error I receive:

ERROR #002100061 has occurred in subroutine SERVE_CLIENTS. |
| Message: |
| The connection between this CFX process and an external process |
| has died unexpectedly. The following error was received: |
| External process : ANSYS |
| Error experienced : Read |
| |
| ** This error is fatal: Terminating ANSYS CFX Solver execution ** |
| Please review the output, error and log files from both the CFX |
| and external processes to determine the cause of the error. |
+--------------------------------------------------------------------+

+--------------------------------------------------------------------+
| An error has occurred in cfx5solve: |
| |
| The ANSYS CFX solver exited with return code 1. No results file |
| has been created.

 flutter April 6, 2010 16:53

Hello Shakeel,

I need more information as to how you set up the problem.

From the geometry side of things, I can tell you that the way my geometry is being set up is by the following steps:

1) I make a (solid) cylinder and extrude it so that it is approximately 5 times larger than the airfoil.

2) I then defined this cylinder as an "enclosure" which automatically makes the program understand that it is a fluid.

Now when setting up your FSI simulation, I actually suppressed everything which is a "solid" (except for the extruded cylinder) and then I generate the fluid mesh. In order to check that you have the correct model set up, when you are defining your problem in CFX-Pre, you should only see 1 3D region.

When generating the solid side of things, how are you defining things?

I find that this is the most tricky part of things in ANSYS 12. What I do is, I define the "solid" problem (supports, FSI boundary, etc) and the quantities I want to solve for. Then, what makes it tricky is that you do not save the file. If you save the file, there will be an error generated because there will be two meshes saved under the right name. What you should try doing is to right mouse click on the "solution" part of simulation and then you generate a ".inp" file and then exit the program. I find that this is the biggest difference between ANSYS v11 and v12.

Anyway, I hope that this helps you and if you have any other questions, please let me know. I hope this helps your error.

gabriella

 Pocket April 6, 2010 22:23

Quote:
 Originally Posted by flutter (Post 253416) I find that this is the most tricky part of things in ANSYS 12. What I do is, I define the "solid" problem (supports, FSI boundary, etc) and the quantities I want to solve for. Then, what makes it tricky is that you do not save the file. If you save the file, there will be an error generated because there will be two meshes saved under the right name. What you should try doing is to right mouse click on the "solution" part of simulation and then you generate a ".inp" file and then exit the program. I find that this is the biggest difference between ANSYS v11 and v12.
Alternatively, if you don't have to fiddle around in the .inp file directly, you can use Workbench and link the structural setup cell to the cfx setup cell.
Then the program directly links your two simulations and creates and links the .inp file directly.

@Shakeel
Your error looks like the mechanical solution encountered a problem.
Look at the ANSYS output file (you can see it in the cfx solver manager) and search for an error in there.

cheers
Daniel

 Shakeel April 7, 2010 09:54

3 Attachment(s)
I have taken a different route to setting up the problem. I have already created a CFD model for the aerofoil using ICEM CFD and have solved it for steady state using CFX. I then wanted to use this CFD model as this will make my work later on in the project a whole load easier. I linked the CFD model with a structural model in ansys workbench (as mentioned by Daniel in the post above, this has limited my options in terms of the meshing but I have done all the meshing) and have set-up all the coupling parameters using CFX-pre. I have also tried to import a structural model but I cannot seem to import an Ansys db file.

The aerofoil is not of standard design so it is difficult to make without a proper CAD software, I was trying to simulate a 2D aerofoil, as I mentioned before I have tried:

1) Solid aerofoil closed section, solves and converges

2) Surfaces over leading and trailing edge solid wing box, works but doesn't converge

3) Surfaces over all the aerofoil open sides, doesn't work at all.

4) I have even tried modelling the thickness and using volume mesh.

Daniel, thank you for your advise, I have found the .out file for the structural model, I also found the following errors:

SURFACE INTERPOLATION MAPPING

PROFILE PRESERVING INTERPOLATION MAPPING
--------------------------------------------------------------------
Sending region name = 1
From sending process = ANSYS
to receiving process = CFX
Transferring variable = DISP

*** ERROR *** CP = 9.922 TIME= 11:58:27
Error in bucket search algorithm for load surface mapping.
Processing complete
--------------------------------------------------------------------

CONSERVATIVE INTERPOLATION MAPPING
--------------------------------------------------------------------
Sending region name = wing
From sending process = CFX
to receiving process = ANSYS
Transferring variable = Total Force
Sending Side: Total Area = 8.16997E+00
Non-Matching Area Fraction = 99.5478%
No. of Un-Mapped Nodes = 784
Receiving Side: Total Area = 2.04224E+00
Non-Matching Area Fraction = 98.2565%
No. of Un-Mapped Nodes = 14882

*** WARNING *** CP = 11.466 TIME= 11:58:29
The component MFLC_1_2_FORC_1 was created for MFLC,1,2,FORC,1 command
to display improperly mapped nodes. Set /debug,,2 for more
information on improperly mapped nodes.

Processing complete
--------------------------------------------------------------------

*** ERROR *** CP = 11.466 TIME= 11:58:29
Error during setup - Skipping the remaining MultiField analysis.

*** ERROR *** CP = 11.466 TIME= 11:58:29
Error during setup - Skipping the remaining MultiField analysis.

NUMBER OF WARNING MESSAGES ENCOUNTERED= 2
NUMBER OF ERROR MESSAGES ENCOUNTERED= 3

This definitely looks like a coupling error with the structural model, but I am not sure how else to model this case. Any suggestions from the either of you? I have attached a few images to show the progression I spoke of above.

Thank you all for your continued assistance,
regards

Shakeel

 flutter April 7, 2010 12:12

A Question

I was wondering if you are trying to do a two way analysis or a one way analysis.

Gabriella

 Shakeel April 7, 2010 13:28

I was hoping to do a two-way analysis.

 Pocket April 7, 2010 22:12

Quote:
 Originally Posted by Shakeel (Post 253562) CONSERVATIVE INTERPOLATION MAPPING -------------------------------------------------------------------- Sending region name = wing From sending process = CFX to receiving process = ANSYS Transferring variable = Total Force Sending Side: Total Area = 8.16997E+00 Non-Matching Area Fraction = 99.5478% No. of Un-Mapped Nodes = 784 Receiving Side: Total Area = 2.04224E+00 Non-Matching Area Fraction = 98.2565% No. of Un-Mapped Nodes = 14882
This (the bold stuff) looks like the source of your error
Are you sure you used the same coordinate system for both the fluid and the solid domain?
Since you build your model in a cad software, can you create the fluid domain in that cad software, too?
If it were in Catia, i would just build the airfoil, then open a new Part, create a Block representing my fluid domain and copy the airfoil(solid) over and boolean substract it from the fluid domain. But make sure that both parts have the same origin and coordinate system. Perhaps choose something important like the 1/3 chord length as the origin.

As for your shell problem, i just read again you use iges for the import. Me and a coworker here at our University found out that iges seems to create instabilities during imports in Ansys 12, step seems to be much more stable.

 Freddie April 14, 2010 12:40

Hello everyone,

i am doing FSI too to study flutter. i am doing the flutter analysis for AGARD 445.6 wing. however my result seem not match the result of previous work from internet. is there anybody can tell me how they calculate flutter speed index. the equation is as follow:

V_f=V_∞/(b_s * ω_α * μ^(1⁄2))

could someone explain it?
and also is there anybody know how to get damping ratio for the flutter from FSI simulation?

 Shakeel April 20, 2010 09:52

Good afternoon everybody,

I have tried what you suggested Daniel, but to no success. Just as a side note that you may find interesting, while conducting the simulation using solid elements I inadvertently made the structural model 400% smaller than the CFD model, however the simulation ran it's course and produced results. This is odd as why would the program allow me to solve a model with a Non-Matching Area Fraction = 25% and not 99.8%.

I am now totally lost as to why the solution will not solve, could it have something to do with the non meshed region within the structural model? Due to time constraints of my work I fear I may have to forget this problem and look for alternative programs for two-way FSI.

Thank you all for your continued help. If I ever solve the issue I will definitely write back.

Shakeel

 Amogh April 29, 2010 13:51

Wing Flutter

Hi,

I am working on wing flutter analysis using fsi in ansys cfx.
I am completely new to cfx. Could anyone please help me with some tutorials specifically related to Wing flutter analysis (using Agard 445.6) using fsi in ansys cfx 12.
Also some initial details like initial boundary conditions applied to the wing and also regarding the modeling of the wing. could it be done as a 2D fsi analysis, if yes please let me know the details.

Amogh

 JohnnyB May 21, 2010 12:12

Flutter in a turbomachine

Hello

I need some advices...i must study the flutter in a turbomachine,I have ansys cfx v12 but i don't know how to start..if someone can help me please send me an email with some advices,maybe an tutorial...
My email adress is :johnnyb_108@yahoo.com

10x

see ya

 Alanseu April 14, 2011 10:23

Quote:
 Originally Posted by Freddie (Post 254613) Hello everyone, i am doing FSI too to study flutter. i am doing the flutter analysis for AGARD 445.6 wing. however my result seem not match the result of previous work from internet. is there anybody can tell me how they calculate flutter speed index. the equation is as follow: V_f=V_∞/(b_s * ω_α * μ^(1⁄2)) could someone explain it? and also is there anybody know how to get damping ratio for the flutter from FSI simulation?
Hi, dear Freddie:
I am a graduate major in aeroelasticity with flutter analysis. I am doing FSI too.
Regarding to the Vf, which is solved under the harmonic motion.
and the damp can be obtained from the F06 file from the results of the simmulation.
The flutter issue is a complicated and interesting issue. As the limited of the File Size, I cannot upload the PPTs for you about the two questions.
If you are still confused in them, you can contact me with alan880702@gmail.com and I can send PPTs for you.
By the way, I have read a lot of the documents about the flutter and simulate some cases to verify my understanding. Could you send me the iges of the AGARD 455.6, I know lots of data about it, but i didnot have geometry of it. Thanks a lot!

 solomedic June 15, 2011 10:19

Agard 445.6 flutter

Hi everybody, All of you people seemed to have worked in somehow flutter analysis,especially Fredie worked on flutter of Agard 445.6. I am doing exactly the same thing in Ansys CFX 13 workbench but the problem I face is that the wing I modeled have diffrent mode shapes modal frequencies than the experimental values,this may be due to missing Geometric properties, Can anybody help me how can I match the modal frequencies to that of experimental values:confused:

 RAHUL MALAREDDI November 6, 2011 10:39

Generating a 3d wing in design modeller

HI
can anybody guide me how to go about doing my project which involves

1. Importing of available grid file of a conventional wing - Onera M6.
2. Putting the coordinates in the software ANSYS.
3. Meshing the wing structure.
4. Analyzing the structure in different flow conditions.
5. Adding a blended wing tip geometry in conventional wing.
6. Carrying out points 3 & 4 above on the resultant wing structure.
7. compare results and establish advantage/ disadvantage.

presently i hv the onera m6 grid file but whenever i try to import it into design modeller it says error in reading file

i hv tried to generate a 2d aerofoil but then too it is showing same error pls help
is there a book to learn how to use ansys 13 tools and troubleshooting

pls help me
Regards
Rahul

 solomedic November 6, 2011 13:15

Hi Rahul, by grid file you mean coordinate? You can import it in a notepad file but therr is a specific syntax that is to be followed. You can write points and it will give you help on how to enter coordinates. Then for meshing you can use the meshing application with in ansys 13. Then you can use cfx as well as fluent for flow analysis. Then the same procedure u can follow for blended wing.
Your questions are quite general,read ansys help and come up with real to the point issue.

Regards

 RAHUL MALAREDDI December 10, 2011 11:03

I tried the help and could generate a wing geometry, however the project can not move further from mesh stage.
While meshing one needs to open model bax connected to geometry in project outline window.
some internal error comes, kindly clarify..:eek:

 solomedic December 14, 2011 00:18

whats the error message?
you have to select material properties such as orthotropic,isotropic, density etc? after that you can mesh it, have you done it? may be your material properties are incorrecct

 hgh March 29, 2016 03:31

Quote:
 Originally Posted by toil (Post 250161) I haven't do any FSI simulation before, but according to my transient computing experience, your time step maybe too long for FSI problem, I do have a mate who was working on FSI simulation, a rectangular wing precisely, his time step is equal to 1E-3 magnitude.
Hello
For rigid airfoil aeroelastic analysis, I do not know what to do to airfoil to move up and down (in terms of time step)?
How do I give the time step?

Thank

All times are GMT -4. The time now is 20:16.