Discussion Groups

Simulation CFD

Reply
Contributor
max.yang
Posts: 15
Registered: ‎08-09-2011

Torque output in 2D rotational analysis

550 Views, 16 Replies
10-25-2011 02:56 PM

I have an issue with the torque output from a 2D rotating surface for a while now and really need to have this figured out as quickly as possible. I wonder if anyone else has the same issue or could shed some light on this.

 

I take a single symmetric airfoil like a NACA0012 and put it in a rotating surface that is circular with the center sitting at the origin (see diagram). I do have a larger air domain around it. I set the material, RPM, mesh and put in a 3-degree-timestamp. The boundary condition is set to no wind speed since I am isolating the torque results. I run it for 4,800 iterations and the torque.csv is created at the end of the run.

 

Over the last full revolution with 30-degree interval, I try to match the hydraulic torque to the hydraulic forces (X and Y) in the csv file, but I find the calculated torque does not agree with the raw hydraulic torque output (see chart). I have accounted for the moment of the airfoil and added that to the calculated torque. The moment of the airfoil is obtained from the wall calculation at each 30-degree interval over a full revolution at the 25% point of the airfoil's chord.

 

I throw the wall calculation torque in the chart just for the sake of curiosity. I am well-aware that getting rotating torque results from the wall calculation is not recommended.

 

This torque issue is observed in both Simulation CFD 2012 and CFdesign 2010.

Please use plain text.
Alumni
marco.mueller
Posts: 35
Registered: ‎08-30-2011

Re: Torque output in 2D rotational analysis

10-26-2011 03:35 AM in reply to: max.yang

Hi,

 

what do you mean with no wind speed? A specified velocity condition of 0? Can you give us an complete overview of the boundary conditions or even attach a share file to your post?

 

Does the same difference of the "force-torque" and torque appear for different mesh sizes?



Marco Müller
Product Support Specialist
Digital Simulation
Autodesk, Inc.
Please use plain text.
Contributor
max.yang
Posts: 15
Registered: ‎08-09-2011

Re: Torque output in 2D rotational analysis

10-26-2011 07:34 AM in reply to: marco.mueller

Marco, attached is the share file with the single airfoil. The scenario has a close-to-zero wind speed of 1e-06 mph in the positive X direction.

 

I was wondering if the hydraulic torque from the csv file is directly related to the hydraulic forces X and Y (or vice versa)? Or are these outputs solved implicitly from different sources?

 

I have not tried different mesh sizes yet, but I will definitely try that now. Thanks for the idea, Marco. I will post my findings once that is done.

Please use plain text.
Alumni
marco.mueller
Posts: 35
Registered: ‎08-30-2011

Re: Torque output in 2D rotational analysis

10-27-2011 01:07 AM in reply to: max.yang

Of course, they should be comparable. But to check this I'd eliminate some error sources. I've found that the rotation region is not perfectly rotating around its center, that's why there occur cyclical issues during the solution:

10-27-2011 9-52-25 AM.png

 

Also the mesh near the interface line can be finer since there is a (small) shift of the values around it:

 

10-27-2011 9-55-30 AM.png

 

A side note: I'd uncheck the output of all quantities that are not relevant for you. Especially the calculation of LMA needs additional ressources (and is anyway not necessary during the solution), so if you uncheck this you will see a tremendous speed-up of the simulation.



Marco Müller
Product Support Specialist
Digital Simulation
Autodesk, Inc.
Please use plain text.
Contributor
max.yang
Posts: 15
Registered: ‎08-09-2011

Re: Torque output in 2D rotational analysis

10-27-2011 02:08 PM in reply to: marco.mueller

Yes, I also observed that strange gap between the rotating and static region. There was no intentional gap created and the rotating region center is placed at 0,0,0 in the CAD file. In the CAD file, the rotating region is just a surface that sit on top of the overall air domain surface.

 

Could there be something wrong that is throwing the rotaional center off?

 

I tried to make another CAD file where the rotating region was "cookie-cut" out from the rest of the region (separated between the regions), but again I saw the gap appearing when the simulation started to run.

 

Any suggestions to remove this unwanted gap and keep the rotating region at 0,0,0?

Please use plain text.
Product Support
Royce_adsk
Posts: 552
Registered: ‎08-24-2011

Re: Torque output in 2D rotational analysis

10-27-2011 02:36 PM in reply to: max.yang

Hi Max,

 

Could you attach your native CAD files? I am not sure why we are not getting a perfect rotation based on what I am looking at.

 

Thanks,

Royce



Royce Abel
Please use plain text.
Contributor
max.yang
Posts: 15
Registered: ‎08-09-2011

Re: Torque output in 2D rotational analysis

10-27-2011 03:00 PM in reply to: Royce_adsk

Sure, this is the CAD file I used in SAT format.

Please use plain text.
Product Support
Royce_adsk
Posts: 552
Registered: ‎08-24-2011

Re: Torque output in 2D rotational analysis

10-27-2011 03:02 PM in reply to: max.yang

What CAD package did you use to generate it?

 

-Royce



Royce Abel
Please use plain text.
Contributor
max.yang
Posts: 15
Registered: ‎08-09-2011

Re: Torque output in 2D rotational analysis

10-27-2011 03:27 PM in reply to: Royce_adsk

I use a CAD software named Shark FX. Do you prefer a parasolid file to work with?

Please use plain text.
Alumni
marco.mueller
Posts: 35
Registered: ‎08-30-2011

Re: Torque output in 2D rotational analysis

10-31-2011 03:17 AM in reply to: max.yang

Can you try to export parasolid or ACIS and then load them into Inventor Fusion which is part of the Simulation package and then start the simulation from Fusion?



Marco Müller
Product Support Specialist
Digital Simulation
Autodesk, Inc.
Please use plain text.