Discussion Groups

## Simulation Mechanical and Multiphysics

Active Contributor
Posts: 43
Registered: ‎05-14-2013

# Non-sense twisting behaviour

158 Views, 4 Replies
05-29-2013 03:46 PM

Hi

I have modelled a foam airfoil with a quarter inch aluminum tube in it's core. I am analysing it's behaviour under torsion. When I subject the foil and tube to a 1Nm surface torque it generates an anticipated twist. However, if i remove the aluminum tube it actually twists less, as shown in the attached images.

The two materials behave as springs in paralel and so this behaviour is non-sensical. Any advice?
The aluminum is much stiffer than the polystyrene and one would expect that adding the tube, all else being equal would reduce the amount of rotation experienced.

I realise that perhaps an MES would be more suited that linear analysis, but for small torques the behaviour remains the same.

Shannon

Distinguished Mentor
Posts: 515
Registered: ‎08-30-2012

# Re: Non-sense twisting behaviour

05-30-2013 07:03 PM in reply to: sapayne

I suggest that you cheat.

Add a few beam elements on the end of the wing, or anywhere that you want to know the rotation (twisting). For example, maybe select the edges on the end of the wing, then select the subentity lines, and duplicate them to a new part. Make the beam cross-section and/or material properties small so that they do not affect the results. When you have the results, you can view the "Results > Contours > Displacement > Rotation > Z".

Otherwise, it is difficult to determine from the two images whether the aluminum tube model rotates more or less than the model without the tube. The displacements are relative to something (not sure what in your model), so there may be something related to where the zero point is located. What I do see in your results is this.

leading edge: displaces up less with the aluminum tube than without the tube.

trailing edge: dislaces down more with the aluminum tube than without the tube.

So the difference in the displacement from the leading edge to the trailing edge may be smaller with the aluminum tube than without.

Be sure to check the reaction forces to make sure they are correct and that the aluminum tube is twisting with the foam. Also check the two rows of "unfixed residuals" in the summary file. This paragraph from the wiki Help explains what I mean:

For static stress analysis, one check on the accuracy is the reaction forces. The end of the summary file (viewable from the tree view in Report environment or by editing the summary file with Notepad) includes a summary of the reaction forces and other parameters. A large unfixed direction residual can indicate a poorly converged solution. Naturally, you can also check in the Results environment whether the sum of the residual forces (the support reactions) matches the applied loads.

John Holtz, PE
Mechanical Engineer
Pittsburgh, PA

16 years experience with Simulation Mechanical
Active Contributor
Posts: 43
Registered: ‎05-14-2013

# Re: Non-sense twisting behaviour

05-31-2013 10:23 AM in reply to: AstroJohnPE

Hi John,

Thanks again for your contributions, very very much appreciated.

I tried your recomendation and I am experiencing the same phenomenon. I have attached another set of photos. Note tht I have switched from modelling the foam airfoil to a foam tube because I can better analyze the results theoretically.

When I twist just the aluminum rod I get the theoretically anticipated angle of deflection of 2* for a 1NM torque. When I add an EPS sheath to the outside and apply the same torque, the angle of twist increases inspite of the added stiffness.
In both cases the tube is cantilevered, so the reference is the base of the tube. There are no large residuals in the summary file.

Thanks again for the help.

Distinguished Mentor
Posts: 515
Registered: ‎08-30-2012

# Re: Non-sense twisting behaviour

05-31-2013 10:54 AM in reply to: sapayne

Hmmmm. Your problem is a mystery.

Please attach an archive of your model so that someone can work with it to figure out what is going on. Other than the obvious mistake of accidentally applying the wrong load (which I'm sure you have checked), I do not have any other hypotheses at the moment.

See the thread "Create, Post, or Provide an Archive of your model".

John Holtz, PE
Mechanical Engineer
Pittsburgh, PA

16 years experience with Simulation Mechanical
Active Contributor
Posts: 43
Registered: ‎05-14-2013

# Re: Non-sense twisting behaviour

05-31-2013 12:22 PM in reply to: AstroJohnPE

I'm having trouble uploading the file, could I email it to you?

Please note that I have left the beam elements as steal bars as I need the face to remain planar during the twist (to determine the neutral axis in the airfoil case).

Thanks again,
Shannon