Rotor damage and side-bands? 1

[Skogsgurra]

Hi,

I'm sitting with this paper machine drive. AC drives. Unruly speed, nothing wrong with drive. Mechies say nothing wrong with machine.

Had a look at motor current. Heavy side-bands (about 15 percent of fundamental). Checked other motors with same load. Side-bands below 1 percent.

Question: Shall I tell them to stop the machine and change motor? Do you think that 15 percent side-bands is a strong enough indication for rotor problems. I will hang if I stop the machine, change motor (250 kW) and it doesn't help. HELP!!

Gunnar Englund

http://www.gke.org

 

[Skogsgurra]

Jraef? Any thoughts?

Gunnar Englund

http://www.gke.org

 

[electricpete]

Have you determined whether the spacing from the line frequency is pole pass?

If yes, as you know a typical limit is about 1%

As a confirmatory check I would look for the slip of this machine compared to slip of other machines. If slip is significantly higher it reinforces the conclusion. Also you may often see oscillating current, increasing vibration and noise over a period of about an hour after startup.

For all the slip calculations you need to know motor speed. If you have a vib spectrum that can be faily easy to pick out. Also a strobe is of course another option.

I believe sometimes dynamic eccentricity (usually enlarged bearing clearances or bowed rotor) can give similar symtpoms). Usually we would look at vib also to see what it shows but it may still be a tough call.

If you have a maintenance availability, there are a few tests you can do in place without disassembly. Single phase test or rotor influence test.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Skogsgurra]

Thanks Pete,

I had to leave site. Machine is still producing. We discussed testing impedance by appying 24 V AC to one motor winding and measure current as rotor is turned stepwise 16x22.5 or 24x15 degrees.

That will probably be done next time it looks real bad. What do you think would be a typical acceptable deviation (max - min impedance) in percent? And what starts to be unacceptable?

The motor is a Siemens 4-pole 690 V 250 kW motor. The folks on site have been told to follow this thread.



Gunnar Englund

http://www.gke.org

 

[electricpete]

Normally you should shoot for 10% voltage. Higher voltage I believe gives a more discriminating test.

Higher voltage of course means faster heatup. No matter what voltage is used, some judgement must be excercized to prevent overheating (which can occur at much less than FLA due to no cooling fan action)

EASA indicates <1% deviation in current is good, >3% is bad, 1-3% is questionable. EPRI states >5% is bad.

You don’t necessarily need to stop to record current. EPRI recommends to turn the rotor continuously at 1/8 to ¼ rpm.

The cycle of current should occur once per pole span so there is not a need to go through a full rotation. But perhaps enough to assure yourself you have seen the repeating pattern. My thinking is you should plan to have power applied no longer than 2 minutes or so to avoid overheating if no RTD's are available for monitoring winding temperature.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Skogsgurra]

Thanks Pete!

Do you think that 24 V on a 690 V motor will heat it? I do not think that there's any risk. Even if single-phasing and "locked rotor". Do you?

The limits are what I really wanted to know. Eng-Tips shows its strength once more.



Gunnar Englund

http://www.gke.org

 

[electricpete]

EPRI recommends 10-20% voltage and EASA recommends 25% voltage. Perhaps that is the reason that the EASA limit is slightly lower.

I would think you want at least 10% voltage (48vac).

I was told it is safe to do the for 2 minutes at 10% voltage so I would assume maybe 8 minutes at 5%.

For a 4-pole motor I would shoot for 1/4 rpm so you could rotate through two pole spans (1/2 revolution) in 2 minutes.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[petronila]

Hello Skogsgurra,
I think if you know the Stator slot numbers and Rotor Bars number you could to have a better interpretation about this vibration cause.
The electric vibration with side bands in electric motor´s rotors could be caused by:

Broken Rotor bars: 1X and modulates at Slip x No.Poles

Loose bars: 1X Possible desbalance effect with thermal sensitivy. Radial. Stator Slot Frecuency Plus Side Bands at +/- no.Poles x Slip.

I think this information could help you: A Summary of AC Induction Motor Monitoring you can visit:

http://WWW.SKFCM.COM/NEWS/APPNOTES/CM3029.PDF

Regards

PETRONILA

 

[jraef]

Gunnar,
Sorry, didn't catch this post in time. It appears as though E-pete and petronilla have given you better advice that I could have generated anyway!

I have always seen 40VAC used for rotor bench tests. I have used 24VAC as a motor winding heater on that size frame, and it takes a very long time to heat the rotor up with it (1 to 2 hours), so I can't imagine it being capable of doing damage on a short test.

Petronila's link didn't work for me, maybe because of the all caps, but this should work. I had that as a reference as well.

http://www.skfcm.com/news/appnotes/cm3029.pdf

I opefully still have somewhere another good paper on using current waveform anaylsis to detect rotor problems. It was in hard copy format, so I need to find it and scan it, but I'll work on that today for you.

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more." Nikola Tesla

Make the best use of Eng-Tips.com
Read the Site Policies at faq731-376

 

[electricpete]

jraef - you are right I was probably way overconservative on those time limits - just a cautios approach if no RTD.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Skogsgurra]

To all,

I just got words that they are still running the machine. I also were informed that the torque was in the 3 to 5 percent range (they hadn't applied any load at all) so the current I measure is mostly magnetizing current. The slip is correspondingly low and the sidebands at +/- 0.15 Hz seem to be reasonable.

I was also told that the motor is a 355 kW one - not 250 kW as said before.

I have three spectra. One from the problem motor, one from a motor without problems, but mechanically connected to the first one and therefore with about the same speed variation, and one from a motor with no problems at all. I can send you the pdf. I think that some of you have my address - send yours and I'll send the spectra for your comments.


Gunnar Englund

http://www.gke.org

 

[petronila]

Thanks jraef for provide US another link. Unfortunatelly I don´t know why the other one doesn´t work.My Apologize to all.
Skogsgurra,
Will be interesting to take a look of the vibration spectra.

Regards

PETRONILA

 

[Skogsgurra]

Petronella,

It isn't vibration. Its is motor current.

Gunnar Englund

http://www.gke.org

 

[electricpete]

At 5% load I would think slip would be so small that distinguishing sidebands would be very difficult and

0.15hz pole pass frequency at 5% load would corresopnd to 3hz pole pass frequency at full load. That would be 1.5hz slip for 2-pole motor or 0.75hz slip for 4-pole motor or 1hz for 6-pole. Those sound pretty high for a large motor, especially if new. (do you happen to know the exact nameplate speed?).

IF slip really is much larger than expected based upon ratioing nameplate slip to power level, that in itself would be an indication of a problem such as rotor bar problem or power supply problem (low or unbalanced voltage). But other possible interpretations could be an error in estimating load, or else this may not be a pole-pass frequency at all.



=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[jraef]

Gunnar,
I found that paper on-line!

Current waveform analysis paper

I'm sending you an email through your website so you can have mine to send that to me. I'd like to see it.

"Our virtues and our failings are inseparable, like force and matter. When they separate, man is no more." Nikola Tesla

Make the best use of Eng-Tips.com
Read the Site Policies at faq731-376

 

[jraef]

Hmmmm... something messed up on that TGML, but it still works. Sorry

 

[Skogsgurra]

Works for me. Thanks.

Gunnar Englund

http://www.gke.org

 

[electricpete]

I think the problem is with the drive or speed control or speed sensing. Some drives have problems controlling speed at low load which goes away at higher speed.

If such is the case, then the sidebands will not be spaced at pole pass frequency.

As discussed above, at this low load 5%, I doubt you can have a pole pass frequency that high 1.5hz. (once again we could help explore that if you have the exact nameplate speed) (and I was already suspicious it was not pole pass frequency as discussed above.... could evaluate a little better if you give exact nameplate speed). This would tend to confirm that you are not seeing pole pass frequency and the problem lies with the drive.

Another way to explore further would be to determine exact machine speed with a handheld strobe or with vibration signature. Then you can compute slip and pole pass frequency anc compare it to your sideband. I suspect this will confirm we are not looking at pole pass frequency and you have ammunition (arguments) to tell your drive people to troubleshoot further.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Skogsgurra]

I have to ask the guys on site. Thanks for the input, Pete.

Gunnar Englund

http://www.gke.org

 

[electricpete]

Now that I think about it, we should not read too much into the fact that this frequency appears too large for a pole pass frequency for a normal motor at this load.

The reason is that slip at a given load is proportional to the equivalent circuit parameter R2. With many rotor bars broken, R2 may appear much larger than it was when the motor was built and slip may be much higher for a given load.

So the bottom line is we can't discriminate further unless we determine actual speed at the time of the measurement.

Knowing speed we can compute slip and pole pass frequency.

Two things would point toward rotor problem.
1 - Slip much greater than expected for the load leve.
2 - Pole pass frequency corresponds to sideband frequency.

or else we may find two things pointing toward drive speed control problem:
1 - slip is as expected for the load level.
2 - pole pas frequency does not correspond to sideband frequency.

Sorry for so much waffling. The bottom line is we can't conclude anything without knowing the speed at time of the test.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.