Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

modeling a torque measurement setup

Please login with a confirmed email address before reporting spam

Dear All,

I am modeling a test setup that is used to measure the torque of a rather unconventional electrical machine called the lateral stator machine. You can see this machine in the attached file.

The torque setup is used to measure the standstill torque of the machine as follows. The stator (which comprises the lateral stator, the windings and the shielding iron, all shown in the attachment) is fixed on a piezoelectric torque sensor. The rotor of the machine is positioned in the stator bore without any connections to the stator (no bearings). The rotor is locked, i.e., it cannot move. When currents flow in the windings around the lateral stator legs, a torque is produced.

Since the stator parts cannot be mounted on a torque sensor directly, they are first put in a plastic chamber, which is also shown in detail in the attached file. The plastic chamber is surrounded by an aluminum housing, which is mounted on the torque sensor. The torque sensor has two rings, and measures the torque applied between them.

With this setup, I have been measuring around 20% less torque than expected. Along with other things, one reason could be that the plastic (the weakest link in there) deforms under torque, hence there is some “lost torque” between the stator and the torque sensor. I want to model this test setup to see if this is really the case.

I believe I will need a 3D model of the stator parts as well as the plastic and the aluminum housings, solved by the Structural mechanics->solid-> stationary solver. Then, I think I can apply the torque on the stator parts as a body load (how do I do that exactly?). But the main question is, how do I define/simulate “what the torque sensor sees”? How can I calculate the torque that is transferred to the torque sensor?

Any inputs will be very much appreciated!

Best regards,
Arda


3 Replies Last Post Oct 28, 2012, 3:12 p.m. EDT
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 28, 2012, 1:57 p.m. EDT
Hi

from your comments, I have one comment, even if your plastic part is deforming, if the full load passes through the sensor there is no reason why the torque should not be measured correctly.

To apply torque the simplest is to use the solid Rigid body connector, but this makes your part very "rigid", another way has been descried in the 3.5a doc and I have dropped a model on the model exchange some years ago.
Basically you set up a cylindrical coordinate around your rotating axis, then you select a boundary perpendicular to the axis, and you apply a phi force of the type

( 2*M*sys2.r/(pi*Radius^4) )

pls check the units and the total value, I might have missed a factor 2 there) where M is the torque sys2.r is the radial distance from the axis, in the cylindrical coordinate system, and Radius is the total radius of your shaft where you apply the torque. You can also replace the pi*Radius^2 by an intop1(1) over the torque boundary

--
Good luck
Ivar
Hi from your comments, I have one comment, even if your plastic part is deforming, if the full load passes through the sensor there is no reason why the torque should not be measured correctly. To apply torque the simplest is to use the solid Rigid body connector, but this makes your part very "rigid", another way has been descried in the 3.5a doc and I have dropped a model on the model exchange some years ago. Basically you set up a cylindrical coordinate around your rotating axis, then you select a boundary perpendicular to the axis, and you apply a phi force of the type ( 2*M*sys2.r/(pi*Radius^4) ) pls check the units and the total value, I might have missed a factor 2 there) where M is the torque sys2.r is the radial distance from the axis, in the cylindrical coordinate system, and Radius is the total radius of your shaft where you apply the torque. You can also replace the pi*Radius^2 by an intop1(1) over the torque boundary -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 28, 2012, 2:07 p.m. EDT
Hi Ivar,

Thank you for your quick reply. That way I should be able to define the applied torque. How can I the measure what is transferred to the sensor mounted to the aluminum case?

About your comment, if the deforming of the plastic is at all a problem, honestly I am not so sure if what I initially said physically makes sense. Would be interesting to see what the model says..

Thanks a lot!

Arda
Hi Ivar, Thank you for your quick reply. That way I should be able to define the applied torque. How can I the measure what is transferred to the sensor mounted to the aluminum case? About your comment, if the deforming of the plastic is at all a problem, honestly I am not so sure if what I initially said physically makes sense. Would be interesting to see what the model says.. Thanks a lot! Arda

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 28, 2012, 3:12 p.m. EDT
Hi

you can calculate the torque over the fixed surfaces, either by adding weak constraints and integrating the lm_... or just by checking the reaction moments calculated by COMSOL.

In any case check the Knowledge base on flus measurements and reaction forces, for advices about mehsing densities, gradients and their resolving.

But be aware, if you have a setup, measuring the torque below and apply ng a torque on the top, this torque will be distributed through the system, if both axes are aligned,the the torque on top should be equal to the torque on the bottom, FEM or not ... (but I might be missing something from your model, as it could be more complex ;)

--
Good luck
Ivar
Hi you can calculate the torque over the fixed surfaces, either by adding weak constraints and integrating the lm_... or just by checking the reaction moments calculated by COMSOL. In any case check the Knowledge base on flus measurements and reaction forces, for advices about mehsing densities, gradients and their resolving. But be aware, if you have a setup, measuring the torque below and apply ng a torque on the top, this torque will be distributed through the system, if both axes are aligned,the the torque on top should be equal to the torque on the bottom, FEM or not ... (but I might be missing something from your model, as it could be more complex ;) -- Good luck Ivar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.