3X Predominant Vibration Peak in 2 Pole Large AC motors 2

[roryrobb]

We continue to experience higher than acceptable overall vibration levels in large, say 900 to 1250 hp two pole, Babbitt bearing electric motors. The predominant peak vibration energy is usually 3X in either the vertical or axial plane.

Current Example: outboard vertical reading 1.3 ips overall, with 1.06 ips at 3X. 2X is .145 and 1X is .725.

What are some of the most common causes of 3X vibration?

 

[electricpete]

Hmmm.

If you view the spectrum on log scale, do you see 6x and 9x rising above the others? I'll bet you don't, but if you do, it implies 3x is a fundamental frequency and you can start looking for coupling with 3 lobes, fan with 3 blades, etc.

More likely, you will not see anything unusual at 6x, 9x etc. Which means you probably have a fundamental frequency of 1x which just happens to have highest harmonic of 3x. It is not unusual to have high harmonics with vibration that high.

Have you attempted a balance?

Have you looked at the bearings? I'll bet they are beat up from the vibration. Inspect beairngs and check clearances.

Does the vibration continue when machine uncoupled?

Does vibration seem load-dependent?

Checked TIR on accessible shaft surfaces?

Prox probe of course provides more sophisticated diagnostics.

Inspected coupling with a strobe? (what kind of coupling by the way... and what does this motor drive?)

Possibilities are wide open at this point.

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

 

[TPL]

We continue to experience higher than acceptable overall vibration levels in large, say 900 to 1250 hp two pole, Babbitt bearing electric motors

Current Example: outboard vertical reading 1.3 ips overall, with 1.06 ips at 3X. 2X is .145 and 1X is .725

Sleeve bearings are frequently monitored with proximity probes and (sorry, no intent to offend) just need to check - are you really measuring in terms of casing velocity with units of ips or are you misquoting displacement measured with prox probes? If so, you might have a glitch or runout issue.

 

[roryrobb]

Thanks guys for your responses. This continues to be a problem for us on motors we are rebuilding with new babbits and proper clearances to shaft and housings. Mostly split endbell types with no way to ensure proper bore alignment to one another other than checking air gaps at each end. TIR is always confirmed to be within .001' to .003" range.

Vibration generally goes away as rotor coasts down. No notable 6 or 9X indications. Doesn't exceed acceptable until almost full speed, I'd say within a couple hundred rpms of synchronous.

We are taking readings with an SKF microlog, motors are not coupled and running at full voltage (4160 and 6900 vac) on a mill table test bed, isolated from ambient structure. Dynamic balance reading with all rotating attachments are all below .1 mils with an IEM 15000 lb. Dynabal.

It just puzzles me that it happens on almost every 2 pole.
Phase analysis is genrally non conclusive of any predominate repeatable pattern.

One note, motors are 60's vintage mfgr dates.

 

[TPL]

So....are the numbers quoted casing velocities in ips or prox probe readings that should be in mils?

If this problem consistently affects more than one rebuilt motor, then it looks like you have a procedural issue.

An overall of 1.3ips on any rebuilt motor would be unacceptable and cause for rejection.

Despite the quote of .1 mil on the balance machine, the resulting 0.725ips @ 1X doesn't suggest a very good balance. Have you had the balance machine calibrated recently?

I would also suggest that TIR of 0.001 or 0.003 might not be good enough - for machines fitted with proximity probes API specs would suggest the TIR (in the form of runout/glitch) should be less than 0.00025inches.

Have you carried out concentricity checks along as much of the rotor as you can access - the journal areas might be round (within the stated TIR) but not concentric with the effective rotating mass.

Rather then just measure TIR, you might want to try and turn the shaft through 10 degree steps and record the dial gauge reading at each step - plot this out to see if you have three lobes on the journal. It is possible that wear in the grinding machine will introduce 3 lobes on the journal and this can translate into 3X vibration.

Have you tried to figure out if the motor structure is resonant at 1X or 3X - try a bump/impact test? It might be wirth trying to get hold of some rundown data (bode or polar plots to get a better understanding of what is happening)