Vibration High Frequency

[nlc]

I have collected spectra of a 90 kw 4 pole 1480 rpm 50 hz supply AC motor with a 12 blade overhung fan mounted direct on the motor output shaft. Of concern is an ampltude of vibration occurring at 1695 hz (approx 68 times running speed of the motor)on all spectra. An amplitude of 6.9 mm/sec rms is the highest at the Non Drive End horizontal.

What could cause this vibration at this frequency??
Can email spectra if required.

 

[DiamondTec]

Email the spectrum and we'll take a look at it, always good to see what you're seeing!

send it to info@diamond-tec.com

 

[Franko]

High frequency could be rotor slot passing frequency, or possibly a harmonic. Noise emanates from cooling ducts, but there also is excitation of the stator and casing. Amplitude typically increases with load. Once we had a very noisy motor driving a plant air compressor we supplied, where number of rotor slots times speed was equal to natural frequency of cantilevered stator projections. Temporary fix was to pack the stator with a special non-conductive damping compound. Re-design had revision to stator projections.

 

[DiamondTec]

Got the spectrum shot in the mail today! More information will help to find a reason for this frequency although as mentioned by Franko slot rate is possible. Let's get some details and think this through...

So far we have an AC motor running at a steady state of 1480 Rpm (50hz power) with no gearbox or such attached. A simple overhung 12 bladed fan direct driven by the electric motor. Digital photos of the equipment also helps...

That should give a 1x rpm frequency of 29.6 Hz.
The fan 12x frequency should be 355.2 Hz.
The spectrum shows a 1692 Hz. or 57.1x motor rpm.

What (if any) other indications are there.. is this frequency detection causing a problem are is it just a non-synchronous peak that you would like to know the cause of?

Did this frequency suddenly appear? After maintenance? Was always there? Is this the first survey of that piece of equipment?

Bearing temperatures are acceptable?

What type bearings are there (i.e. SKF 6313) on the motor or shafts?
Spectrum appears to have multiple harmonics of the 1692 Hz with sidebands, what order are the sidebands (ie 1x. .5x) at?

If your software lets you, can you place a harmonic cursor on the 1692 Hz peak and "divide" your cursor by different amounts (by 2, 3, 4, 5 etc.). That may help find the actual "source" peak if the 1692 Hz is actually a high harmonic that is being detected. Once you get the "divide by" cursor you should see all the peaks line up. You may find it is a bearing fault frequency around 4.7x (139.1 Hz.) and the 12th harmonic is showing as the most detectable peak.

If you can email a zoom of the 1692 peak and also a zoom of the lower range showing 1x orders etc. that would be helpful also.

In any vibration analysis you can't provide TOO much data.

Oh, if you have the capability to monitor the vibration in "real time" mode, you can help determine an electrical vibration from a mechanical vibration by watching the 1692 Hz peak while turning power off to the motor. If the peak immediately goes away during coast down it is a good indication that it was an electrical vibration peak.

Other obvious checks are that the motor does not have "soft foot" or that the motor case is not being deflected somehow. All mounting hardware tight, balance is OK etc.

 

[DiamondTec]

Whoa son! Hey, ignore some of that last post from DiamondTec. It's my fault, I was answering generic questions from another cubicle and didn't know someone here was going to post a reply (we all have access to this site with the same login name).

I read the post and found some errors (duh?)

Ignore the frequency stuff, we work with "orders normally" (not frequency) and I misunderstood a question yelled across the room about changing from 60hz to 50 hz power.

I looked at the spectrum you sent and it does look like the sidebands are at 4x orders that may indicate an electrical vibration. Monitoring the waveform data during shut down you may help determine electrical verses mechanical vibration problems.

If you send the other data talked about I'll look it over (myself).

Again, sorry for any confusion.. I need to pay attention to what (and why) someone asks me something.

 

[DiamondTec]

Hi Rod,

I got the background and other spectrums, etc. and will check them out. Things are hopping around here and I have to look at these kind of things "off the clock" as it is so I'll try to get back to you Thursday.