New phenomenon in Vibrations of Electric Generator 2

[INDAR]

Hi guys!
A new phenomenon has been seen in a Run Down of an electric generator. It is shown in the attached graphic.

The machine has ball bearings with grease in two sides. One side is fixed and another one is non fixed in axial direction. The horizontal line is appeared only in 15% of manufactured machines. Why?

The colour-map generated is obtained in a Run Down test. So there is no electric supply. The machine goes to 2000rpm and then the electric supply is switched off. Now, the rotor speed goes down until 0rpm. The horizontal line appears near 1600rpm.

Do you need more data? Wich the reason is that generates this horizontal line?

Thanks in advance.

 

[SomptingGuy]

Looks like 1/2 order knocking (impacts) to me. Can you hear anything unusual at that speed? I can't begin to guess why it's 1/2 order though.

- Steve

 

[INDAR]

Hi Steve,

At that speed the noise is normal. There is no impact noises. The orders are the diagonal lines that you can see in the colour map. The horizontal line isn't a order. I don't know what it's...

Personally, I think some non linearity (but wich?? maybe the ball bearings??) is producing that horizontal line. I don't know...

The next fb-waterfall shows the same phenomenom but in this case to 850rpm.

Thanks in advance,
Best Regards

 

[SomptingGuy]

I can read colormaps (I write software that creates them!). My point was that at your problem speed, 1/2 order excitation is evident, which isn't something I'd expect to see for a simple rotational machine like this, hence my suspicion that this is connected to the problem. Horizontal lines in colormaps ususlly point to something misbehaving (knocking, instability, non-linearity, etc) so that all harmonics are excited.

- Steve

 

[MotorVib]

Do you know what the rotor naturial Freq?

Also how big is this unit (kw), operating speed?

I have seen things like this before, and it was related to a naturial freq. I would do some more diging to see if this is repeatable. If this is a naturial freq, it can be anything in the system.

Chris

"In this house, we obey the laws of thermodynamics." Homer Simpson

 

[electricpete]

I agree with Somptingguy in terms of the frequency content of that yellow line. Looks like 1/2x (12hz) with harmonics of 1/2x (12hz) occuring at machine speed ~ 24hz. The fundamental frequency of motion appears 12hz, but it is not sinusoidal (presence of harmonics)

And I agree with Motorvib - has the characateristics of a resonance. If you look when the machine speed gets down to 12x (the lower part of the graph), you see a slight intensification of the yellow again - which supports the idea that 12hz is a resonant frequency.

Putting it together, as 1/2 decreases through resonance, the magnitude gets so high that it excites non-linear effects, giving harmonics.

Another interesting thing - there is a faint vertical line at this frequency and it's harmonics. Then as we move down the chart, the left-most vertical line (12hz) tends to curve left toward slightly lower frequency and next right veritcal line (24-hz) curves slightly right.

It is a little bit of strange behavior. Thinking out loud - two things come to mind that may or may not be of interest (see for example Den Hartog or Harris' Shock and Vib Handbook)

1 - Self-excited vibrations of rotors. Typically they occur at rotor resonant frequency. Also typically during increasing of machine speed (a different scenario than your test), we would not any of these instability effects until rotor speed gets near/above twice the natural frequency. That is generally the threshhold speed. As rotor speed continues above that, the vibration would continue at the natural frequency (not the rotor speed). Reasons for this type of behavior include: oil whip... internal hysteresis of the rotor due to loose parts such as loose laminations... fluid trapped in rotor... etc. Although typically we don't think of this type of vibration continuing as rotor speed decreases below twice the natural frequency).
2 - If I'm remembering correctly (and I'm not sure I am) I think he talks about criticals at half speed due to the effect of gravity.



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

 

[Tmoose]

" criticals at half speed due to the effect of gravity" .

One Hartog's example (might have been others) was for a rotor of non-uniform bending stiffness.
2X4 wood beam, 2 pole motor rotor, gets twice per rev vertical bounce on earth with shaft ~ hrizontal

 

[electricpete]

Thanks Tmoose. Twice per rev would of course be different. Maybe I was mistaken. I'll see if I can find it.

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

 

[GregLocock]

Am I right in reading it as rpm up the y axis and Hz across the bottom?



Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[electricpete]

That's the way I read it. (reading points off the 1x line, the vertical axis of rms is around 60* higher than the horiziontal axis of hz).

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

 

[electricpete]

Let's try that again (correction in bold)
That's the way I read it. (reading points off the 1x line, the vertical axis of rpm is around 60* higher than the horiziontal axis of hz).

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

 

[SomptingGuy]

If you squint you can just about read Hz on the X-axis and rpm on the Y-axis.

- Steve

 

[INDAR]

Hi guys!
Sorry about the bad quality of colour map. In this one:
x-> Hz
y-> rpm
colour scale-> vibration level

In the 2D graphic:
x-> rpm
y->Vibration Level

In some cases, the vibration level is important and the machines should be rejected.

The natural frequency is 30% far from nominal speed (1200rpm) by design.

One question: When you say 'criticals at half speed due to the effect of gravity'... This is only for flexible rotors or rigid too?

Finally, I have found a paper of Dra. Muszynska that says that the horizontal line is due to Nonlinear Stiffness. . But it doesn't says what causes that phenomenon.. It says:"Horizontal line is not an indicator of crack. This is due to Nonlinear Stiffness".

See soon!
Best Regards,

 

[INDAR]

Hi guys!

I am thinking about this problem . I would like to know your opinion about the next one:

Is it possible that a difference among diameters of inner balls of ball bearings produces this nonlinear stiffness showed as horizontal line?

In negative case, how could you see this another problem?

Best Regards,

 

[Tmoose]

"Is it possible that a difference among diameters of inner balls of ball bearings produces this nonlinear stiffness showed as horizontal line?"

Have you gotten familiar with the rotational frequencies of components and events in a ball bearing? If one or a group of undersized or oversized ball significantly changed stiffness, that influence would be rotating at about 0.4X shaft speed.

 

[Tmoose]

Pete said "Twice per rev would of course be different."

I think you had it right.
The frequency of excitation for the [2 pole motor - 2X4 lumber] effect is 2 vertical bounces per shaft rotation.

If INDAR's calculated rotor "natural frequency is 30% far from nominal speed (1200rpm) by design", AND that natural frequency is first bending mode, AND the natural frequency is above rotating speed, THEN discounting damping and what not, if he has a two pole motor then a "critical speed" will be apparent as the rotor coasts thru about 65% nominal speed. The frequency of the vibration during the critical however will 2X the rotor rotating frequency.

(The fact a single calculated/designed rotor natural frequency is mentioned suggests bearing and support stiffness may not have been included in the cyphering process)

 

[electricpete]

Wow. I think you may be on to something there.

Best guesstimate of the frequenies where the horizontal line appears from attached is 1800 rpm (30hz) machine speed and 12 hz vibration frequency. (note I positioned the vertical red dashed line based on the underneath vertical yellow line... that vertical yellow line easy to see in the original version but not so much in this revised version).

13hz/30hz = 0.43

Which as was mentioned is a very typical value for Fundamental Train Frequency - the frequency that the cage and rotating elements orbit around the bearing. (Note in my simple mind I think about the motion of earth about the sun... FTF equates to one year and BSF equates to one day). If you have one larger rolling element, it would show up at the same position (relative to a stationary reference frame) at frequency FTF.

Do you have a bearing part number? Manufacturers and vib programs provide fault frequencies based on bearing part number.

Putting together the story line - we have a resonance at 13hz, and a forcing function (FTF) which gets to 13hz when machine speed is 1200 rpm. And perhaps the vibration magnitude is so high that it excites non-linear effects giving those harmonics of 13hz.

But here's a part that doesn't fit that particular story line - why don't see see the 13hz vibration again due to 1x excitation (unbalance etc) when the machine speed passes through 780 rpm?

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

 

[electricpete]

Sorry, forgot to include attachment which just shows projection lines to determine the frequencies

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

 

[SomptingGuy]

Note INDAR's second post:

"The next fb-waterfall shows the same phenomenom but in this case to 850rpm"

That's close to 780 rpm. It has to be bearing-based.

- Steve

 

[SomptingGuy]

INDAR,

Thanks for this thread. It's making me realise how much I miss the detective work that follows some relatively simple vibration measurements. It's all a big puzzle and the measurements have the answer.


- Steve