I have an interesting analysis question. I have 4 AC motors driving boiler feed pumps that all have indications of possible cracked\broken rotor bars. I find it very hard to believe that I have 4 motors all next to each other with broken rotor bars. Axials look great. What else should I be considering here? I was told that the motors seemed to be oversized for the application and we are running about 20 rpm over nameplate. Thanks in advance for any input you might have. I have attached a few of the high res spectrums.

relweb.doc (122 Kb, 112 downloads) rotor bars

Actually the spectrums you show doesn't have enough resolution to see broken or cracked rotor bars. Its very hard to see yours. you need a 50 hz spectrum with 3200 lines to look for sidebands of about number of poles X slip i think it is. How are you running faster than name plate is the motor on an inverter ?? Reguardless of load a standard induction motor across the line shouldn't even run at name plate slightly lower. Am I right here.

These are 12,000 cpm, 1600 lines, I would think that is high enough resolution. Thanks for your reply.

Assuming you have verified the sidebands are spaced at pole-pass, the next logical step would be to perform current signature analysis. If you get high pole pass sidebands around 1x in the current spectrum, that is a stronger indicator of rotor bar problem than vibration. There is also additinoal info that may be gained looking up around 5*LF and looking up around RBPF in the current spectrum.

Dynamic eccentricity (bent rotor or orbiting at 1x within clearances) can also in theory cause vibration pole pass sidebands. You could check for bent rotor with a dial indicator on accessible portion of shaft extension if motor can be shutdown. Also could monitor with prox probes but first/easiest check is current signature analysis.

I would say it is not out of the question that 4 similar machines of the same vintage, same manufacturer, same operating history could develop rotor bar dagradation at the same time. 2-pole motors are the most at risk for rotor problems. Other risk factors would be high inertia (usually not applicable to pumps), frequent starting, and age (degradation occurs over the course of many starts).

true I had 2 2500.4160 that failed within 6 months of each other. Did your problem appear after a while or was it there first time u took data.

Actually, a motor can certainly run 20 RPM above "nameplate" if nameplate is, for example, less than 1780. Seems to indicate these motors are lightly loaded, which is significant if you truly do have a bars issue.
MCA is key to accurately diagnosing rotor bars. If possible, check these motors as close to full load as can be done.
Something to keep in mind: Some "barred" rotors are designed to run loose in order to allow for thermal growth and some cast rotors have natural voids that can simulate high resistance connections.
Good luck.

I think you may have side band space = 2LF
What is the temperature at the frame of motor Is it 90-100 degree C? what class of insullation?

Check for the temperature and insullation test.

Good Luck
Phoo

I am not sure I am understanging what you are saying about your load. You say you are running about 20 rpm over your nameplate rating. Your spectrum is marked at 3547 rpm. What I am hearing (reading) is your full load rpm is around 3525 rpm. The boiler feed pump motors I have seen do not have that kind of load slip. What is the full load rated rpm? When you say over do you mean that your current speed is less than your nameplate speed? That would tell me you are overloaded and not oversized. But I do agree that the spacing does look like rotor bars. I have seen many BFP motors of the same vintage and design (15) start to fail within a few months of one another. You have some good advice in the above replies.

Nameplate speed is 3520 rpm. I am very confident that they are sidebands of pole pass freq. This is the first time we collected data on these motors. We will do it again in 12 months(not my choice)! I am recommending current signature analysis.

Can you clear something up for me on these sidebands' "definition"? Some are saying "pole pass" and some are saying "rotor bar pass". One reply does mention slip times number of poles.

Which are they or are the names given the same? Has the sideband name been changed in the last few years?

Or are they slip frequency times the number of poles? Looks as if the difference from 1X to the sidebands varies with the speed on all the spectrums, so seems like their position would have to be related to some "speed" or "slip" times something, huh?

Like BFP3, the slip is 52.5, so are the sidebands at 103?
BFP5 slip is 45 so are they at 90?
BFP4 slip is 59.5 = 109?

I am definitly no electrical person so I am at a lose at what everyone is talking about here if they (sidebands) are not "slip X poles".

Ralph,

I was also confused about the reference to rotor bar passing frequency.

I'm sure that what Doc is looking at is pole pass frequency or the number of poles times the slip frequency and that your interpretation of the sideband spacing is correct. Maybe Doc can post one of the spectra with the sidebands shown to confirm it.

I would immediately go to where the book points here too, but on 4 motors at the same time?

Doc,

I didn't know you had a 1 year period on this. That would make the decision easy for me. I would hit them all with a red flag and say that you cannot guarantee that they will still be running after the 1 year period is up.

I alway consider (and let my clients know up front) that any period over 3 months is vibration analysis and not predictive maintenance. Based on the data that you have to go one, are you willing to say that any of these motors will be around in September, 2007?

I would report what the data has told you. There may be problems with the rotor bars, they could be significant and they shouldn't expect to be able to wait a year before checking them more thoroughly. If they aren't willing to check it out further, they should be prepared to replace them all very quickly.

I'm sure these people are reasonable. After all, they do work for the United States government.

Danny

Here is a zoom of one of the spectrums. I recommended that they perform Motor current signature analysis. Thank you all for your input and ideas, it was all greatly appreciated.

Sorry_about_the_confusion.doc (46 Kb, 60 downloads) rb's

quote:

Sorry about the confusion. Here is a zoom of one of the spectrums. Even at 3x’s I have sidebands of 84.75 which is pole pass frequency. Slip * # of poles.

The above is a quote from Mike Doc's document.

I just want to make sure from now on that I "know" what one is speaking of when the term is used, so::::

Is "pole pass frequency" the correct terminology when refferring to these sidebands, when one is talking about bad rotor bars?

Normally your main rotor bar defect frequency peak will be number of rotor bars X RPM. Most will be way out at 79-100 CPM also with side bands at 2X Line Frequency, or 7200, your program band cursor will confrim that, you will also normally have a peak lower down in the spectrum smack on 7200 cpm confrming the 2X Line Freq. Your spectrum even looks similar to loose leads, and stator problems. You definately need to extend your FMAX out to 150-200,000cpm

That is the formula for spotting rotor bar freq. if you do not know the number of rotor bars in a motor.As the rotor bars get worse, your side bands will also increase past the alarm. Time to shut down!

I've been on a hard drive persuing these problems myself for the past two years, ever since I had a 250hp Toshiba blow apart on me one night...Due to broken rotor bars and the extemely high 1X associated with that. Once your rotor bar(s) start to go bad you lose that circular magnetic field which maintains the circular motion of the shaft. It becomes an eccentric or eliptical rotation which will destroy a motor.

Good Luck!

Rod

VENDOR ALERT

Mike,

Not sure if you have had dealings with PdMA in the past but you can contact them to help and if your lucky you can get them to come do demo on these machines.

The PdMA Regional Mgr for your area is Mr. Harold Joyce. His office number is (800) 476-6463 x-122 or his cell is (813) 240-5607.

You can go to www.pdma.com for info on what they can do.

Regards,
Marty Wilkins

Rotor Bars problems will often produce 2 x Line freq. sidebands around Frb = rotorbar pass frequency = # rotor bars. Most motors will produce these frequencies and the fact that they are there is not a problem. If these sidebands start to change then that is an indication of a potential problem. Thes are typically 50 to 100 time running speed.

As the problem becomes more severe I have observed that you start to get Fp = pole pass frequency = #pole X slip sidebands around running speed and running speed harmonics.

In this case the indicated spectrum shows the second indication

The motor is a 3600 2 pole motor.

Running speed is 3557

Slip = 3600 - 3557 = 43

Fp = Slip X # Pole = 43 X 2 = 86