Torque: Understanding Motor & Pump Shaft Misalignment

[Ask1122]

Hi all, a few days ago our lecturer has told us something which i do not really understand, and i was hoping if someone can clear up the issue for me.

The following is what was said: we have a motor and we have a pump, the pump and motor shaft is joined by a coupling. If there is a misalignment between the pump and the motor shaft, the 'TORQUE' will causes the pump shaft to bend. Now this is what i don't understand, i always think of torque as a twisting force, how can a twisting force 'bend' something? In order to 'bend' the pump shaft, wouldn't you require a 'bending force' (which i thought it was the force exerted by the bearings which causes the pump shaft to bend when it is mis-aligned)? I think i am missing a basic concept on torque here, it will be really helpful if someone can clarify for me!

And secondly, if the pump and motor shaft is misaligned, wouldn't there be bending shafts for both the motor AND the pump? Why is it only the pump shaft that will be bend??

Sorry for my elementary questions, but i just can't seem to grasp this concept well. Thank you for the helps guys!

 

[CFDFEAGURU]

It would seem that your prof is trying to tell you that misalignment of the shaft would cause the shaft become eccentric and not revolve in a circular manner but instead revolve in a slightly elliptical manner. All shafts can handle a slight amount of eccentricity because the shaft is not "perfectly balanced". The eccentricity will cause additional torque to be applied to the shaft and the coupling. Your prof is calling the eccentricty "bending" of the shaft.

If the eccentricity is great the shaft can be damaged. Sometimes this is referred to as "whipping" of the shaft.

Thanks
Matt

 

[Q_Goest]

Hi ask. Welcome to the board.

Unfortunately the way the question is proposed, it doesn't make much sense. The torque obviously doesn't create a bending moment in the shaft. The misalignment does. The torque only causes the shaft to rotate such that the stress in the shaft also rotates with respect to the shaft.

 

[FredGarvin]

Perhaps it is just a bad choice in words. Maybe if the prof had used the word LOAD instead of torque.

 

[Mech_Engineer]

You are imagining torque (moment) as a load that must always be applied axially to the shaft, which is not the case. You should instead think of a moment as a vector, where the center axis of the moment can be pointing in any direction (to apply pure torsion, a moment's vector just happens to be colinear with the shaft's axis).

Take for example a simple beam bending problem, there are many cases where a moment is applied to a beam causing bending.

http://www.efunda.com/formulae/solid_mechanics/beams/casestudy_display.cfm?case=cantilever_endmoment

In the case of a torsional moment that is being applied to a shaft, if the moment's vector is slightly out of line with the shaft it will cause both torsion and bending in the shaft.

 

[FredGarvin]

ME,
I thought about that for a little while, but I couldn't reconcile in my own mind how a coupling arrangement with an offset would apply a torque to the end. To me it still seems like it is a distributed load along the length of engagement of the coupler.

 

[Mech_Engineer]

ME,
I thought about that for a little while, but I couldn't reconcile in my own mind how a coupling arrangement with an offset would apply a torque to the end. To me it still seems like it is a distributed load along the length of engagement of the coupler.

Well, I was imagining a situation where a moment that is being appied to a shaft is off set not only in position but also angle. So take for example a moment that is 1 degree out of parallel with the shaft (no offset), then it seems to me that you would split that moment into a torsional component and bending component (something like 3-D force vector addition/separation using trigonometry).

If the moment's vector is parallel but offset from the shaft, then the resultant load on the shaft would be have to be split into a moment and force at the end of the shaft for it to be statically solved; the force would cause bending in the shaft while the moment would cause torsion.

 

[Ask1122]

Thanks for the reply guys! Just wanted to find out this though, so how will the shaft get damaged by the misalignment? I thought that the misalignment will create a bending force at the bearing (because it will be leaning more to one side when it spins), is this assumption even close? Thanks

 

[Mech_Engineer]

Thanks for the reply guys! Just wanted to find out this though, so how will the shaft get damaged by the misalignment? I thought that the misalignment will create a bending force at the bearing (because it will be leaning more to one side when it spins), is this assumption even close? Thanks

Generally that is true. But, rather than try and design the system so that both shafts are perfectly aligned, it is smart to instead use a coupling which allows for angular and/or radial misalignment between the two shafts. The design and application of the coupling will depend on your particular requirements for power transmission and misalignment. Examples:

A helical flexible coupling:
http://heli-cal.com/Library/website-pictures/Aclampmap1.jpg
http://heli-cal.com/cm/Products/Flexible-Couplings/Home.html

or U-joint style coupling:

 

[Ask1122]

Thank you very much! This really helps! :D

 

[FredGarvin]

Usually the assumed loading is just a radial load on the bearing, not really a moment. I can't imagine a normal installation in that the misalignment would be great enough to bend a pump or motor shaft. Even if one aligns them by sight, you would be close enough to not do that much damage. You would not be doing justice to the coupler or bearings, but bend a shaft? That's a lot of load. Perhaps with very large pump/motors this is indeed an issue.

 

[FredGarvin]

Well, I was imagining a situation where a moment that is being appied to a shaft is off set not only in position but also angle. So take for example a moment that is 1 degree out of parallel with the shaft (no offset), then it seems to me that you would split that moment into a torsional component and bending component (something like 3-D force vector addition/separation using trigonometry).

If the moment's vector is parallel but offset from the shaft, then the resultant load on the shaft would be have to be split into a moment and force at the end of the shaft for it to be statically solved; the force would cause bending in the shaft while the moment would cause torsion.

I see where you're coming from. I think that would indeed be a worst case scenario.

 

[Ranger Mike]

having sold laser alignment systems for machine tools and pumps I can tell you that all kinds of bad things happen when the motor and the driven component ( pump, drill, precision way, etc..) are not aligned. All things driven have bearings. there bearing have a degree of tolerance to " float or self center" to a finite degree..but..once that point is exceeded, mucho friction and $$$ repair occurs..hence the coupler noted above, to permit a slight amount of misalignment. don't forget...the plant maintenance guys are still using technology that was used to build the pyramids..string and spirit levels...in some cases advance dial indicators to CENTER the shafts but..the angle of the coupled shafts still must be aligned and this is the benefit of the laser alignment system..nuff of my sales job
improper alignment cost in dollars bug time..takes more energy to turn a misalign shaft..one example..turbines used to make electricity ..these are 100 feet long monsters...when drive is shut off and the turbine is properly aligned it takes an hour for the shaft to stop spinning,,when misaligned it stops in 10 minutes...

 

[CFDFEAGURU]

stops in 10 minutes...

That is lot of additional force on that shaft. WOW.

Thanks
Matt