Rotor Whirl after dropping onto ball bearings at 22,500 rpm

[nanobot29]

I'm fairly new at rotordynamics and I'm having some trouble interpreting some test data. i would appreciate some input. We're working on a motor that runs on magnetic bearings. The mag bearings lost power during a test and the rotor dropped onto a set of auxiliary ball bearings while spinning at 22,500 rpm (375 Hz). There's a 0.010 inch radial gap between the ball bearing inner race and the rotor O.D. Test data from an accelerometer sitting in the endbell shows a peak-to-peak frequency of 210 Hz after the drop. Is this frequency a result of the rotor whirling? my co-workers are convinced that this 210 Hz is the whirl frequency. I have trouble believing this because this would generate an enourmous whirl force.

could it be possible that the rotor was whirling at half the frequency (105 Hz) and this whirl frequency may have excited a mode in the housing or endbell which has a natural frequency of twice that amount? is it possible to excite a mode while applying a force at a frequency of half it's natural frequency?

The reason i dont believe that 210 Hz is a whirl frequency is because a previous test showed that the rotor whirled at 40 Hz during a drop at 18,000 rpm. huge difference. unfortunately, our data aqcuisition tools were turned off during the recent test and all we have to go by is accelerometer data from the EndBell.... which shows 210 Hz peak-to-peak.

Thanks for any info or insight you may have.

 

[GregLocock]

"is it possible to excite a mode while applying a force at a frequency of half it's natural frequency? "

In a linear system, no. Real systems aren't linear, so maybe.

I can't make much sense of the rest of your post. The whirl frequency I know about is directly measurable with a hammer and an accelerometer as it is the bending mode of the rotor.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[nanobot29]

Greg, thanks for the input. So whirl can only occur at the rotor's bending mode frquency? can any of the rigid body modes induce whirl? the rotor's bending mode is at 500 Hz

 

[GregLocock]

No RBMs can't cause whirl, they are at 0 Hz

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Tmoose]

Can you post the time wave form?
is it possible to rig up some prox probes to track what the shaft is doing?

Are the shaft and race ID lubricated in any way?

A rotating shaft mounted to ball bearings with clearance between inner race and shaft tries to drive the race with friction.
Kind of like this.

http://www.slotcardigest.com/motori...gears_technical_information/internal_gear.jpg

http://www.tangs.com.sg/Files/Hula Hoop_Art_com.jpg

If the friction was perfect and the (horizontal) shaft heavy enough to lie at 6 oclock the shaft would spin at one frequency, and the race would spin (creep) a little slower at the ratio of shaft/bore diameters.
If there was immense unbalance so the heavy spot pressed the shaft hard against the race bore (similar to if the inner race was locked to the shaft with a setscrew), then the shaft and race would spin at the same rpm, but with the rotating assembly eccentric by 0.005 inch with corresponding unbalance.

Or, have enough inertia to make a loop without any support near the top, as at 1:51 here -

http://www.youtube.com/watch?v=eYDjV7ajuMo

More likely some intermediate condition would occur, with the shaft clawing its way up the rotating race from 6 oclock, then slipping down, kind of like a surfer changing position on the face of a wave.

http://www.youtube.com/watch?v=MaCSXX1KxH4

I imagine there would be plenty of undamped energy available to lock on to any convenient resonance of the rotating assembly and/or structure.

 

[nanobot29]

The attached images show vibration with respect to time. the data was taken from X and Y accelerometers attached to the bearing housing.

after looking closely at the data, i now feel that the 210 Hz peaks are not caused by whirl. Im thinking that maybe the impact force excited a rigid body mode. The data shows that the bearing fractured at 0.8 sec du to the 40 G load.

Image 1 - moment of impact
Image 2 - 0.7 sec after impact
Image 3 - 0.8 sec after impact - bearing failure