Not entirely sure what to make of this and I hate being in this situation.

I just baselined 3 new high speed blowers driven by 500 HP 6900 VAC induction motors supported by sleeve bearings. The uncoupled data looked good, but when I coupled the unit up I ran into an anomaly. The variable diffuser was opened at intervals throughout the testing so I got data at different motor loads up until we hit close to 98% amps. As load increased, a modulating noise not unlike what you hear with broken rotor bars began to get mreo and more evident. The data showed an increase in 2xTS (this is a 2-pole motor) separated from 2xLF. In addition, there were numerous pole pass sidebands around running speed and out to several harmonics. The current imbalance as read from the MCC controller was about 2%. Unfiltered overalls on the motor remained less than 0.07 IPS and RBPF was slight at best.
This same pattern was exhibited on all 3 motors. Now I know these motors are brand new and my client was skeptical there was even a problem. To be honest, I was nervous about it as well. I mean, we're not talking about a blatant bearing defect or misalignment, right?
So the question is: Do I pursue getting motor current data to verify this is not a rotor bar problem? Would the sleeve bearings have anything to do with it?
Also, I once heard from a guy who knows a guy, etc, that some barred rotors are purposely made with a little looseness to allow for thermal growth. This would certainly simulate broken rotor bars right after startup, no? Any validity to this?

Thanks in advance, all!

Do you have any data to support your question? This is not a $5 call you are talking about.

Thanks for your reply, Dave. And no, this is not a $5 call, good guess. .

I have attached for your viewing pleasure a low freq hi res reading taken on the motor DE bearing, vertical. The first spectrum is 0-400 hz, the second is a zoom around 1xts and 120 hz. Those sidebands are at pole pass freq. The third spectrum is the same as the second, only in log scale.

For the record, I have not condemned the motors but have recommended further data collection (motor current).

Like I said, I hate these scenarios when all 3 of a particular machine are behaving in the same odd way. Thanks again for the reply and let me know if you need anything else.

Frank

RBMware_Document.pdf (44 Kb, 41 downloads) Data

Frank,

If these are Reliance Motors, they may indeed have "loose rotor bars" for a while, well, for ever really. Seems a ROckwell engineer decided that since there was so much trouble keeping rotor bars tight, he would let thermal growth do the trick. They are loose when cold, supposedly tighten up when up to operating speed. I ran into a bunch of them being used for pumps. The RBF was there, all the time. It was much more amplitude when cold, but it was always there, more than usual motor. If I remember correctly, they had radial bearings though. I don't know about any having sleeve bearings.
I don't remember what the style number was, but Ron Brook may be along after while and can tell you. He is still with ROckwell.

Good luck.

Dave

Why do you know there is no misalilgnment and what about run-out data? What kind of couplings? Were there acceptance data folders sent with the new motors? On a wild thought note; did you collect data on the terminal side? Sorry for the questions but this is not like being there and it's early and no coffee.

We are prime manufacture of medium to large size motors in India. We have never heard, anybody leaving any gap between rotor slot & rotor bars. Rather they are kept tight by using thin steel liners in between, so that bars are tight & donot vibratte while running.

DKSONI

These are Siemens motors. The motor and blower were aligned together at the factory and verified again once installed on site. Would misalignment on a sleeve bearing machine result in pole pass sidebands?

The coupling is a flexible shim type. Acceptance data sent from the factory was, of course, acceptable. All they used, though, were the installed prox probes, no FFT. Interestingly enough, I checked out some of the prox probes readings on the motor as it came up to full load and some of them were up over 2 mils and oscillating slightly. For a 2-pole motor, isn't that a little high?

No, I did not collect data on the terminal side, Sam, but the other side. No problem, grab yourself a cup of grim

Frank,

Dynamic eccentricity will cause pole pass (PPF) SBs to appear (as broken rotor bars will do). So, a possible scenario of events may be as follows: misalignment -> shaft shifts off center and orbits -> dynamic eccentricity -> PPF SBs.

I would've verified shaft movement if possible.

David

Huh, I had thought dynamic eccentricity would only cause PP SB around 2xLF and rotor bars causes PP SB around running speed and harmonics. Am I wrong on this? Also, will very slight misalignment of this unit actually cause the motor shaft to run more eccentric than it already is, seeing as how these are sleeve bearings and they naturally run slightly off center?
How would you have verified shaft movement, by looking at the prox probe data?

Thanks

1) You state that you have low amplitudes @ 1XRBPF how about 2XRBPF what are the amplitudes? The noise you speak of, is it modulating? Just at certain loads? Does any HVA reading show signs of looseness? Pole pass frequency should no be here, new or old motor. Another test for bad rotor bars is to perform an online current analysis thru CT and PT. I have also used old needle style clamp amp, watching for the current to flucuate rythmically. If there was a serve rotor bar issue, the motor typically will run hot and will not achieve nameplate design. With a machine offline you could do a rotor influence check with motor assembled, to look for cracked or broken rotor bars (could be low influence rotor) You should involve/consult the factory with the data you have.
2) Can you provide H V A data on each motor plus put all three motors HVA in a waterfall plot together?
3) Notice the variance in amplitude from 1X to 2X in your data. (1x .01 ips 2x .08ish ips) Does this same 8:1 ratio show in each axis.
4) Notice at the base of your peak @ 2x it has a small mound of energy like a "hump", 1 x does not show this. This mound of energy and ratio difference can come from a resonance, thus explaining higher 2x amplitude

1) 1xRBPF is very low with an amplitude ranging from 0.003 to 0.008 IPS, 2xRBPF is not any more prevalent. There is a definite increase, albeit slight, in the number of 120 Hz sidebands around 1xRBPF as load increases. The noise coming from the motors is definitely "modulating" in time with pole pass s/b (I timed them). Noise increased with load. No real signs of mechanical looseness in the sense that the noise floor does not become raised and there are just a few low level harmonics. I agree with your mention of doing MCA, I have made that recommendation to the client.

2) Yes, see attached.

3) No, not really, see attached.

4) I think that is legit, not resonance: it increases with load.

Thanks for the insight thus far....

Frank

RBMware_Document.pdf (123 Kb, 16 downloads)

quote:

Also, I once heard from a guy who knows a guy, etc, that some barred rotors are purposely made with a little looseness to allow for thermal growth. This would certainly simulate broken rotor bars right after startup, no? Any validity to this?

I would say there can be element of truth to the discussion of loose bars. As identified below, there have been many strategies for swaging rotor bars and some of them involve securing the bars in the center but not swaging towards the ends so that the rotor bars can expand evenly to address the stress which would otherwise be caused by differential expansion (copper expands more than iron due to higher temperature and higher coefficient of expansion.)

Excerpt from EASA Motor Failure Root Cause Analysis Handbook

"Different manufacturers have various philosophies about
rotor bar swaging. Some of those ideas have also changed
with experience, so a manufacturer using a tight cage
design today may have built rotors with a loose cage design
in years previous. This information is provided to illustrate
the complexity of the question "How tight should rotor bars
be?"
One manufacturer swages only one end of each bar.
Another drills and spot-welds the mid-point of several bars
to locate an otherwise loose cage design. Still others use
loose cage design, but use a centering ring between
shaft and endring to prevent endring and cage distortion.
Still another actually stakes the laminated core, driving
lamination edges into the bars, to prevent movement.
The preferred method is to swage each bar uniformly, at
even intervals, to tighten the bars in the rotor. One manufacturer
reported in an IEEE paper that the expected life of a loose cage design was 4000 starts."

I don't the "loose" rotor bars would create any false rotor bar symptom since the end rings should also move with the rotor bars and there should be no loss of contact between the end rings and the bars. If contact is lost, that's a problem which can deteriorate further and eventually lead to failure.