I would like to see if there is anyone who may have an idea about what is going on here.

2500hp 300rpm 180 rotor bars and 216 stator slots. (Fresh rewind)

The main peak is at near 60400 with sidebands of 2x LF and the harmonic of 60400 also has sb's of 2xLF I have attached some spectral data for your review. There are amplitudes of .7ips on the stator.

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

Is the noise a concern for people or for machine health?

Are you saying the noise increased, or the machine is just noisy?
If it has always been that way, I wouldn't worry about machine health.

In my experience, it is not uncommon for slow speed motors to be fractional slot motors and it is not uncommon for fractional slot design motors to be very noisy. (Both families of large slow speed motors at our site are fractional slot and noisey). However, based on your data, it's not the case for your motor.
At 300rpm,60hz that makes it 24 poles.
Slots per pole-phase group = 216slots/(3phases*24poles) = 3
So you apparently don't have a fractional slot winding. A fractional slot would be a fractional number of slots per pole phase group, rather than an integer.

The noise certainly can still be electromagnetic since there are a wide variety of contributing factors that can increase electromagnetic noise.

High motor noise can also be from windage effects and perhaps a few other causes.

The noise isn't as much of a concern as the vibration levels with the high frequency peaks and harmonics. The noise just made us more aware of the issue. This increase with the new refurb.

Jason,

It may not help in elimination of the noise source but just in order to confirm or rule out that noise is at 60400 CPM, why not hook up a regular microphone to the voltage input of your analyzer?

You did not post the spectra.

David

sorry about that.

Reading_collected_on_stator_iron_near_ODE.doc (120 Kb, 64 downloads)

What are the amplitude 'units' on the plots? I can't tell.

quote:

2500hp 300rpm 180 rotor bars and 216 stator slots. (Fresh rewind)

Isn't the 60430 CPM the stator slotpass? Based on the order in the spectrum. I know it says 218.17 orders, but maybe your speed is off just a little. Figures 280 RPMS.

Is it possible the noise is a point on the rotor passing the slots causing a "wind" noise? Were the peaks there before the rebuild of the motor and the noise was not?

Maybe the airgap is off or the motor has a softfoot or something and the casing/stator is warped.

This must be an open motor for you to get on the stator iron as described in the plots.

may i know is internal blowing or something similar is normally included in the motor rewinding?

we had 1 case where kind of similar frequency generated with unusual noise when running. the motor team performed internal inspection and it was found out that one of the rotor ring having circumfrential groove. and it was believe that the welding slag is the culprit as we found lots of loose welding slag at the upper part of the motor winding.

Here are some answers to the questions.
The amplitudes are in ips.

The speed is 295. The 60400 is 2xLF higher then RBPF and 1xLF lower then slator slot pass.

Yes it is an open motor (vertical)

I have thought about airgap but the motor shop said everything was well within specs when they sent it out and with the large size (10.5" dia)and vertical I don't think an installation could cause that? Is it possible this is a sideband of one or the other frequencies and it is showing up with harmonics?

That's the first time I have seen (what appears to be) a harmonic of a sideband. That is not expected in general.

I would say it seems possible that a harmonic of a sideband can occur in presence of looseness. Imagine you have resonant amplification of lower sideband so the whole motion is almost sinusoidal at that lower sideband frequency. Now imagine at one limit of travel within that sinusoid there is an impact from looseness. You have harmonics of the impacting frequency = harmonics of that sideband. But I have never seen it happen.

Another thought - is it possible you had the wrong number of rotor bars recorded?

quote:

The speed is 295. The 60400 is 2xLF higher then RBPF and 1xLF lower then slator slot pass.

Hard for me to deal in cpm.

E-Pete,
180 Rbars, That was my first question as well and the motor shop went back and verified this information on the sister motor they currently have.

Ralph,

Big numbers or small numbers all the same to me. I am just used to cpm. Divide by 60 and let me know if the smaller numbers change anything. Sometimes when I translate frequencies from hz to cpm problems seem easier for me to understand. Once again though the speed is 295cpm (4.9hz) and the stator slots are 216 and rotor bars are 180. Frequency in ? is 60400cpm (1006 hz)

If you take 60,400 cpm and subtract 7200 cpm to get 53,200 cpm. Now divide by 295 rpm and get 180 that equals the # of Rotor Bars. There are situatons where the sideband has higher amplitude than the center frequency. El'Pete can probably do some math magic on this!

Walt

I have heard that sidebands on one side of the frequency compared to the other side can mean something. Has anyone else heard this before? Has anyone ever noticed the sidebands change sides with change in rotational direction of the motor?

I just reaching here but if questions don't get asked answers will never appear.

Taylor’s gear analysis Handbook may have an explanation, on page 45, he explain how amplitude modulation translates into FFT with USSD (upper side, side band) and LSSD (lower side, side band). In your case the carrier is the rotor bar as mention by Walt and it is hyper modulated (side bands higher than the carrier) by the line frequency (120 Hz). It translates by one stator slot that sees every rotor bars passing and modulates the signal by the line frequency. In gear theory looseness (page 100) is translated in FFT by a higher LSSD modulation of the gear mesh. If that theory could apply to motor than I would say that one conductor on a stator slot is loose in its channel and vibrates at 120 Hz. It is probably a little off so its “sees” the rotor bars passing at motor’s speed. If it is the case than you are ok until that conductor touch the rotor and short.
Best regard, Marcel

Sideband higher than the center frequency: there area lot of ways to look at it.

As was mentioned above, if we have simple complete modulation (multiplication), then the center frequency disappears and all that is left is the sideband.
cos(wc*t) * cos(we*t) = 0.5 * [cos(<wc-we>*t) + cos(<we-wc>*t)]
where wc is carrier and we is envelope. Note on the rhs we have the sidebands above and below wc, but no center frequency wc remains.

What if one sideband is much higher than the other - could be result of more resonant amplification at one frequency than the other.

With all that said, I'm not sure that modulation analysis applies directly to this situation of RBPF +/- 2*LF sidebands. Or if it does, it is certainly more complex than just imagining that the normal 2*LF is modulated/multipled by the passage of rotor bars at RBPF. The "simplest" analysis I have seen is given by Alger "The Magnetic Noise of Polyphase Induction Motors" from 1954 IEEE documents. It shows that under load, the effect of rotor and stator slotting is to create three flux waves:
1: cos((R-S-2P)*theta+2*Pi*<RBPF-2LF>*t)
2: cos((R-S)*theta-2*Pi*<RBPF>*t)
3: cos((R-S+2P)*theta+2*Pi*<RBPF+2LF>*t)
where R is number of rotor slots and S is number of stator slots and P is number of poles and theta is the mechanical angle coordinate and 2LF is twice line frequency and RBPF is rotor bar pass frequency.

The first of the flux waves is the spectral LSB (RBPF-2LF), the second is the center frequency, and the third is the USB (rbpf + 2LSB).

They each have a different spatial "mode shape" which is described by R-S-2P, R-S or R-S+2P. In general, the cylindrical stator is stiffer against high mode numbers (which cause a lot of bending at tight radiuses) and more flexible to low mode numbers which cause gentle bending. So IF R > S, THEN R-S is positive and the first one (RBPF-2LF) has the lowest mode number where the stator is likely most flexible and will likely create the highest noise/vibration.

These waves may rotate forward or backward. If the sign of the coefficient of theta is the same as the sign of the coefficient of t, then they rotate backwards (opposite the direction of the rotor), if opposite sign they rotate forwards (same direction of the rotor). Depending on the values of R, S, and P any of these might end up being forward or backward waves.

I would think that reversing direction of the motor by swapping two input leads should reverse everything (the direction of the rotor and the direction of each of these harmonic waves). So you end up with kind of a mirror image of where you started. I wouldn't think that would change the pattern. But that is just me. I will be the first to admit that these harmonic flux patterns in motors are much more complex than I can understand.

To return to an earlier part of the discussion - the higher sideband is not a surprise, that seems very normal to me. The harmonic of the sideband (if that's what it is) is very surprising to me (as it was to Jason). I can't understand the reason for that.

For this discussion 120 Hz is not to be considered a 60 Hz harmononic, but rather a fundamental frequency caused by magnetostriction. I have measured sound around power transformers and generators. The sound spectrum contains 120 Hz and harmonics of it at 240, 360, and 480 Hz and not multiples of line frequency at 60 Hz.

Walt

Jason highlighted peaks at 60300 and 120600. They appear to fit the following pattern:
1X = 295 cpm
RBPF = 180 * 295 = 53100
USB = RBPF + 2LF = 53100 + 7200 = 60300
2*USB = 2*60300 = 120600

The very last frequency above is the unexpected one. It is the one I was calling a "Harmonic of a sideband". But instead of calling it a harmonic of the sideband (2*(RBPF+2LF)), maybe it is better to view it as a second sideband above 2*RBPF (i.e. 2*RBPF + 4LF).

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

Just an update;

The motor shop has the sister motor in their shop with very similar vibration signatures and is recommending a core replacement (basically because all other avenues have been ruled out, but this is not a guaranteed fix). Has anyone heard of this causing this type of vibration?