Greetings

We have a 2-pole 1750 HP sleeve bearing Louis Allis motor that has gotten the best of us since late 2004. This motor drives a Howden screw refrigeration compressor that is crucial to our process.

Timeline: About 2 years ago, this was the best running machine. One month prior to a planned compressor PM and the sending out of the motor for cleaning / refurbishing, a stator temperature alarm prompted motor vibration analysis. No significant vibration problems were seen, but I did note a new 3X axial vibration peak at .09 ips. During the repair, a new drop out shim pack type coupling was installed to facilitate maintenance. The unloaded coupled machine was tested in Dec. 2004. Axial motor vibration revealed the same 3X peak, but now at .15 ips. Days later while under full load, the axial motor vibration at 3X increased to .52 ips and the motor prox probe vibration due to 1X increased to 4.00 mils. The danger was set to 4.50 mils. The coupling was focused on. The coupling fit and balance was looked at and adjusted. The motor base and motor feet were looked at and machined. Hot alignment checks followed and the 3X axial motor vibration stabilized at around .20 ips.

In April 2006, the motor prox probes increased to 5.40 mils vibration at 1X. In June 2006, an electrical failure at the motor occurred. The repair shop had to re-lead but not rewind the stator. The motor interior was dirty. The coupling side bearing revealed wear and some cracking, but was repaired prior to our inspection. Baseline solo testing revealed in spec vibration, but a 3X was observed in the axial direction. This 3X dropped out immediately when the power was cut.

During the coupled test in July at 93 percent load, the initial vibration looked promising. But after only 10 minutes, the motor prox probes trended into alarm as high as 7.00 mils. Motor axial vibration increased from .20 ips to .45 ips. We had taken high-resolution spectra at the horizontal positions to look for rotor and stator problems, but never at the axial positions. Non-averaged data was watched in real time at the axial 3X peak. It was seen to modulate up and down. A high-resolution spectrum of the 3X revealed pole pass frequency side bands as high as .14 ips. The motor was sent out and 3 broken rotor bars were found. Testing by the repair facility prior to repair revealed abnormal shaft growth when power was applied. The rotor was re-bared. Retesting of shaft after repair revealed no abnormal growth.

Motor sent local shop to test after repair. Overall vibration was acceptable at less than .08 ips during the cold start run. But within 10 minutes the axial overall increased to .11 ips due to a 3X. The bulk of the 3X energy remained after the power was cut.

The baseline coupled test of the motor at first revealed .16 ips axial vibration and below alert prox probe vibration. The motor was cold and had not been run recently. Within 8 minutes, the axial vibration increased to .40 ips and high motor prox probe vibration at 6.50 mils was recorded. After hot alignment check and changes, the axial vibration shot up to .60 ips and motor prox readings remained above the danger level. The machine had been down for a couple of hours, but was apparently still warm from the prior 3-hour test. The axial vibration was at the same 3X peak. High-resolution spectrum of the 3X revealed no rotor bar type vibration. CONCLUSION: No satisfaction, only disappointment to date. What might be causing our 3X motor axial vibration?

Nice problem description for an interesting case. I suggest a bump test of the motor end bells to rule out or rule in resonance as a contributor to the vibration. It definitely does not sound like resonance is the root cause, but it may be that one of the motor end bells has a natural frequency near 3X, which would help explain the axial vibration. To get a believable bump test, I recommend uncoupling the motor and rolling it by hand just enough to get it on an oil film. Then when you hit the end bell, the shaft will not add stiffness in the axial direction. I had a group of motors ring out at almost exactly 180 Hz one time.

If resonance is present, maybe the end bell is cracked. Maybe there is is design problem. Whatever the cause, you would then have to stiffen to move the natural.

If you could find the source of the 3X axial, then the 1X radial might be a balance/thermal problem that only shows up at full speed and load. In this case, if the vector is repeatable, balancing might result in vibration you could live with.

Michael Titone

G Man,

Sure sounds like the rotor is still bowing as load is applied. There is often a cross coupling effect at the bearings between the radial and axial vibration.

The fact that the axial vibration shows up at 3X is interesting, normally it appears at 1X. Is the 3X vibration only apparent on the spectrum or is it also on the waveform? If its not showing on the waveform, does the waveform have a sawtooth pattern?

Is it possible for you to track the 1X, 2X, and 3X amplitude and phase during startup and loading? Are you getting any axial vibration on the driven equipment?

Thanks for the feedback to date. The reality of the situation is that ops does not want to spend any more money on this problem. The B motor has been pulled for cleaning and will be put in the “C” slot of this problem motor. Due to piping and locations, the “C” compressor is the most efficient. A new motor purchase, waiting game, and new base are now planned for the open slot. I will share the feedback to date with maintenance and see if there is interest to pursue the comments received here.

The axial waveform does indeed show a 3X pattern that is saw tooth in appearance. The compressor has no such frequency or axial vibration of concern. As far as phase, I could track the phase from the Bently and observe the frequency spectrum, but I no longer think that ops would support that without a good reason.

The prox waveform is not a perfect sinusoidal pattern at 1X, but it is close to it. The handheld radial vibration at the motor is not a problem. During the max vibration as captured at the radial motor positions, I did capture 1X harmonics up to 6X that were .04 to .08 ips in amplitude. The waveform for this one revealed 1X spaced impacts, per my interpretation that were 10 Gs peak to peak. Thanks again to everyone for the feedback to date.

G Man,

Definitely an interesting problem. Couple of questions I would like to clarify. It appears from your description that increase in axial 3x is accompanied with increase of 1x only in radial. Is that correct?

If so, it would be useful to live monitor phase Vert vs. Hor on the IB bearing and also accross the coupling. This may indicate developing bow as motor warms up. Of course, as mentioned above, 3x in axial may be close to axial resonance.

I also wonder what is happening with axial readings on the OB bearing?

Please keep us posted.

David

The axial OB and IB vibration is equally bad. From a cold start up, vibration is livable. But after 8 minutes or so, the axial vibration was observed to increase and by then the Bently prox probe monitor at the motor radials begins to go into alarm one after another.

Prior to having the motor re-bared, we did capture phase data across the motor through the Bently system while the compressor was loaded. The phase and 1X mils vibration for the X and Y probes follow. OB X 195˚ at 5.4 mils, Y 307˚ at 5.8 mils. IB X 331˚ at 4.4 mils, Y 184˚ at 6.38 mils. Keep in mind that prior to the rotor being re-bared, the repair shop measured abnormal shaft growth with applied power. As a result 3 broken rotor bars were found and the rotor was re-bared. Testing afterwards revealed no abnormal shaft growth per the repair shop. I failed to capture additional phase data during this last test, mostly because of our perceived failure of solving our problem and knowing that further actions to troubleshoot would be limited. I messed up. ……

My vote would also be that the be rotor problem is not fully corrected. Mechanical looseness of the bars may be creating harmonics which are magnified by resonance at 3x. Change over time of course consistent with thermal rotor bow. There is always tension when rotor heats due to more thermal expansion of copper than iron. If the bars are uniformly gripped by the rotor core, this is an even tension all the way around the rotor and the final result is only axial tension in the rotor core with no bending. If some bars slide within the core to change that pattern, bending can now occur.

Another thing to note of course you're bearings won't last long once you exceed pk/pk shaft displacement of 50% bearing clearance so there may be some bearing damage as well.

My one and only experience with rotor bar looseness on a large motor was also Siemens which as I understand was somehow connected to Louis Allis. If it's any comfort to you, that problem kicked our butts as well. After much on-site troubleshooting including a consultant and an OEM rep, the motor went finally back to the Norwood factory who told us the only solution was to replace the rotor. It seems the L-shaped rotor bar design supported at 3 points cannot be satisfactorily be reswaedged. My understanding is the 1970's and early 80's vintage from that OEM had a less-than-great rotor bar/support design.

This message has been edited. Last edited by: electricpete,

Did you have torsional resonance checked before coupling was replaced?

Thanks again to all for the support and comments. I'm just a vibration tech, but I know that we did not do a torsional resonance check before the coupling was replaced.

In fact, I am not familiar with that check. In reference to the coupling, I have always questioned the new coupling because our problems started after both units were rebuilt and when we switched to the new coupling type. But, we have documented 3X axial issues on this motor while on solo on numerous times. These solo checks with the 3X include vibration testing at the motor repair shop.

One additional fact, three identical units were installed in the late 90s or so I believe. The “A” motor never did run right from the beginning and the supplier eventually replaced it with a Siemens of a different frame design. The problem then was thought to be rotor related.

New motors are 6 month delivery. Repair shop was not cheap.

C_waveforms.xls (38 Kb, 13 downloads) excel

Gman,

Have you inspected the coupling during operation with a tunable strobe? Look for distortion in the shim packs.

This type of coupling is notoriously intolerant of improper axial positioning.

Good Luck,

G Man

The TWF prior to hot alignment rather looks like one in a bump test and is excited every rev. Possibly, something is ringing at a low frequency. Although it appears that alignment still left 3x in place. BTW, you have never posted the spectra.

The coupling looked normal when viewed with a strobe light. We did see axial 3X vibration at the motor shop while on solo. Data was less, but present. See attachment for vibration data. Thanks for all replies to date.

C_MTR_AXIAL_vibration_Sept_testing_06.xls (194 Kb, 13 downloads) Excel with vibration data

Does the fact that 3x & 1x pattern for seismic vibration is quite different from that of prox probe vibration tell us that there is a possibility of structural resonance at 3x?

If not, why are the patterns so different?