Hello

We have got a 11 KV, 50 MW, 600 RPM, 50 HZ Vertical Hydro Salient Pole Generator.

Non-contact proximity probes (4-20 ma output) are fixed on the individual bearing housings and the following are vibration readings in microns

UGB – Upper Guide Bearing LGB – Lower Guide Bearing TGB – Turbine Guide Bearing

Condition ---------------------------- UGB ----------- LGB ------- TGB
Unexcited, 600 RPM (100% speed)--- 253 / 241 ----- 300 / 312--- 182 / 201
Unexcited, 870 RPM (145% speed)--- 317 / 258 ----- 312 / 352-- >500 / 344
Excited, 600 RPM (100% speed)----- 247 / 235 ----- 318 / 310 --- 210 / 165

Are these vibrations due to generator unbalance ? Or, due to misalignment ?

The screen shots of the vibrations and tabulated vibration readings (I could not get the tabulation straight into the post, sorry) are also attached.

Thanks in advance for your help.

Muthukumaran

This message has been edited. Last edited by: edison123,

Vibrations_Screen_shot.pdf (638 Kb, 38 downloads)

It looks like you are asking to diagnose this with just overall vibration. Overall can be the starting point to say you have a problem, but it rarely suffices to determine the cause.

Thanks Bill.

These readings are actual displacements measured by the proximity probes. So aren't they at 1xrpm ? Pls correct me if I am wrong.

If they are regular probes having 4-20mA output, there is normally no tracking filter so it´s rectified broadband I would guess but you never know. If you "guess" the 1xRPM dominate it could be a case of "electrical unbalance" where the irregular magnetic force move the center of rotation possibly. In large on site built hydrogenerators it´s not uncommon to balance for the electrical influence to have it run as good as possible when on full load where it should be. If this a normal generator, assuming the above and the behaviour has not been such before, consider some winding problem, maybe, possibly, considering all if´s and but´s or I may be wrong. Olov

Thanks Olov.

The generator is 2500 km away from where I am sitting. It was recently refurbished with new stator and rotor windings. The rotor poles were weighed (weight difference of about 10 kgs in a pole weighing 2500 kgs) and poles were distributed evenly.

The first two vibrations are without excitation (ruling out electrical problems) at 100% and 145% speed. The third one in excited but at no-load.

The generator operator feels there is no difference before and after refurbishment. But the CONSULTANT says it is unbalance without providing any basis for his claim. This is the first time the proximity probes are used (as a part of refurbishment). Now the consultant is imagining ghosts.

Would it help if I posted the signature here ?

This message has been edited. Last edited by: edison123,

Pls see attached the signatures as seen by DCS.

Vibration_signatures.pdf (388 Kb, 37 downloads)

I'm with Bill. Olov has good lines of thought. But my take just from overally with fundamental problems mostly at 1X & 2X would graphically from a hard plot line through centers call it as misalignment but there isn't real bias for this. YOu really need a full analyses and your outside consultant should be able to provide this by gathering thier own data - have them plug-in to the prox probes???!!!...

edison, You only have prox probe readings?
Send him a vibrometer with FFT to take absolute readings of the bearings so you don´t chase glitches on the shafts and that you see what frequency it is. If your machine is of that size it´s not abnormal to do a balancing at full operating speed, no excitation if required, and then at full load if required. If you are lucky it will be operational both at run up and operation full load or you need to figure out a compromise. As I am like I am I would use absolute probe (velocity :-) )for that input to do the balancing. Olov

Edison, from our experience monitoring / balancing hydros of similar size to yours, we would only use proximity probes as the bearings are relatively very stiff. We would always plug our analyser into the buffered output of the installed monitoring rack so that we can see all of the components including any changes in the DC gap value as the machine excites as this can indicate magnetic centre errors. However you are severely restricting your options trying to use 4-20ma signals as these only give you the overall values and no information as the the vibration components. We have seen hydros with high 2.3 to 2.8 x running speed vibration which was a resonance excited by the runner band seals. To help with your decision making you could try watching the vibration as the machines run down/ run up. If purely balance the vibration should be related to the square of the speed. If there is a dominent alignment issue the vibration level can stay high as the speed drops. With out the component information you are in the dark and I would be very cautious of making any preictions.

Regards

MDE

Thanks OLI & MDE. FFT it is then.

MDE - You mentioned "If purely balance the vibration should be related to the square of the speed." Exactly, my point. Here we had 145% speed (which squares up to two times) and yet the generator vibrations went up hardly by 5 to 25%.

Hi Edison, I would advise not just FFT, but 1x and 2x amp and phase. Record both during the run up and when excited. Recording the phase is very important. If we look at an extreme case where all the vibration is at 1x. Unexcited 150µm at 90˚ excited 150µm at 270˚. The overall value will show 150µm in both states but the balance vector has changed by 300µm. Be also aware that particularly in vertical hydros some situations can mask others. We had a vertical frances machine with 400µm p-p at 1x. Was assumed to be balance but did not respond to weight change. Was in fact the thrust collar slightly off square on the shaft by 1/1000th on an inch. This caused the shaft to "cone". The other comment I would make the comment that the readings you supplied particularly the lower guide are quite high and it would definitely be good to get them down in the long term.

Regards

MDE

Thanks MDE. Do you happen to know the vibration limits for this speed ?

Muthukumaran

Ok. I got hold of DIN ISO 7919-5 - Evlauation of machine vibrations by measurements of rotating shafts.

It says, for 600 RPM, vibration limit for long-term operation is about 130 microns peak and 240 microns peak-to-peak.

Do the proximity probes read peak or peak-to-peak ?

Attached is the Trendline of the unit indicating speed and various proximity probe readings (Microns)

My observations are

As the speed is increased/decreased

1. LGB Y (yellow) and LGB X (orange) rose with the speed initially up to 100% speed, then stayed constant during the speed change to 100 to 50%, 50% to 80%. Then, from 80% to 100% speed increase, both these values rose by about 7%.

2. UGB Y (rose) was crazy from 0 - 100% - 50% - half the duration of 80%. From next half of the duration of 80% to 100% and then to 0%, stayed with the speed.

3. UGB X (red) stays with the speed.

4. TGB Y (pink) was crazy from 0 - 100% - 50% - start of 80%. From 80% - 100% - 80%, it stayed constant at 180 - 220 microns. From 80% - 0%, it went crazy again.

5. TGB X (brown) generally stayed with the speed (except for some crazy behavior at around 1510 hrs). Then from 80% - 0%, it went up.

My questions,

A. Does the time waveform indicate any generator dynamic unbalance ?

B. Why are some readings going crazy ?

This message has been edited. Last edited by: edison123,

2007-08-30_WCU3_heat_run_-_vibrations.pdf (147 Kb, 15 downloads)

I don't see what I call a time waveform. These are trend plots of amplitudes.

Ok, Bill. I changed it to trendline.

Any conclusions from it ?

Is it not possible to get items like time waveforms, spectra, or coastdown plots?

I agree with Bill, almost impossible to to draw any conclusions from this data other than to question the pink data as this looks very suspect.

Thanks Bill and MDE.

We have planned a spectrum analysis today. Will get back to you with the results.

Site balacning was done today and results are enclosed.

The vibrations at rated speed, unexcited have come down

UGB X/Y - 92/110

LGB X/Y - 183/185

TGB X/Y - 191/130

Thank you all for your tips.

Vibration_at_100%_speed_-_Unexcited_-_after_balancing.pdf (149 Kb, 0 downloads)