Hello all!
I've recently seen some "strange" behavior on one of our Reactor Coolant Pumps. The RCP is a large vertical pump driven by a 9000 horse power vertical motor. Picture of the pump shaft area is in the attached document. The pump shaft is monitored by two proximity probes 90 degrees apart looking at the pump half coupling just above the seal. Proximity probes are also installed below the lower motor bearing and above the upper motor bearing.
The vibration is "spiking" periodically on both pump shaft probes (see plots). The lower motor probes are also showing a change in vibration when the spiking occurs, but it is very small. The change in vibration is all at 1x with a shift in phase angle of ~10 to 15 degrees. Vibration will trend up for ~20 minutes then slowly subside back to normal values.
Orbits and spectra don't show anything except and increase in 1x vibration during the excursion. I think this may be a slight rub, maybe in the seal? The seal is a 3 stange N9000 series. I've never seen or heard of this type of behavior on an RCP and was hoping someone else had?
Thanks,
Ed Hudson

P-32D-vibespiking3.pdf (333 Kb, 45 downloads) Vibe spiking

Hi Ed,
Used to work with similar babies around 1989 or so it was the deluxe with grating floor to stand on when working around the coupling, we trimbalanced them every year before startup with "washers" in the coupling bolts during warm up, one shot only, to reduce bearing contact in motor bearings, still most times you could find contact areas in the bearings. Since we run 50Hz with minimal design change I guess. We had some additional features. They sure would produce weird orbits but we didn´t have so much monitoring back then, part from vector of 1 and 2xRPM and the orbit scope. So the only thing I remember to have seen is touching in the motor bearings and loose fit of the big flywheel. Balancing we did basically compensated for the repositioning of the flywheel during revision and the mode shape allowed weights in the coupling to compensate for that to some extent as levels was reduced on the complete machine. That is also something that is weird with this one. I would have expected things happening at the other probes also since they are close and it´s heavily connected. Only thing I can think of is if there is a glitch on the shaft that comes in and out of target for that probe set. I can´t imagine that it would be possible in this design. Have you looked at the raw signal when it happens? Is it "sinusoidal"? Sorry not to have any good solution. Olov

Eyeballing the overall "envelope" of this behaviour (amplitude and phase)it seems to rise and fall in an almost half-sine manner over some days. I'd be looking for a process parameter that relates tightly to the vibration behaviour.
Is the primary loop pressure rock steady?
Can you review seal parameters during this period?

I'm somewhat familiar with those motors but not at all with the pumps. One cause of intermittent behavior on RCP's that I've heard of (and you probably aware of as well) is shifting of the flywheel (when the 3 3-piece key upgrade is not installed). This usually gives step change and should appear on motor, so doesn't sound relevant.

That's just a longshot I wanted to mention. You are a lot more familiar than me, and based on that I'll bet you're on the right track about the seals.

The real reason that motivated me to respond to this post was to mention some references and more specificaly mention an author :

EPRI TR108480 - Considerations for Vib Monitoring Diagnostics of RCP.

EPRI NP6337 - Vibration Monitoring of Reactor Coolant Pumps - Guideline and Reference Data - by Dr. Ron Hartlen

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

Have not seen this vibration behavior on our RCPs. They seem to be real events...that is, they are corroborated on several channels. That assumes it is not some kind of power transient in your Bently communication processors? I think that could be discounted by observing normal looking behavior on other channels sharing the same communication processors. I would also be interested in observing a spectra during the event, to be sure that the behavior still appears to be entirely synchronous, rather than broadbanded noise.

If it is a real event, I sure agree with Ron...that is to try to corroborate the vibration event with some process event. My first guess would, as Ron already suggested, be to look at seal parameters. Our seal return flows are recorded in the Control Room. I believe return temperatures are available deep in some hard drive as well. The frequency of the events look as though they may be roughly coincident with the frequency that your operators may be making-up to your Volume Control Tank. This is a manual activity that the reactor operator performs approx one to five times per shift to control coolant system chemistry. A change in volume control tank level could change backpressure against seal return, affecting its flow. A change in volume control tank temperature due to the frequency of make-up could also affect seal RCP seal injection temperature.

How would something like a seal parameter affect RCP vibration...I dunno? I suspect, however, thermal affects could change internal clearances...or return flow could affect the ability of the seal faces to mate properly...somehow affecting shaft vibration.

I'd sure be interested to hear any outcome.

If you suspect cracking problems verify that the tacho pulse versus eddy probe output phase is stable. Only shaft crack I have seen proven data from verified tracking of a crack for fairly long time (days-weeks) and there was a continous phase shift for the 2xRPM until shutdown and then the shaft broke during coast down but that was a turbine. Our RCP design was very close to shaft resonance at full operation due to that the design did not take in account the lowering of resonance induced by flywheel fitting getting pretty loose at full operation thereby giving relatively high unbalance sensitivity. This was at the time verified by calibrated FEM analysis and a lot of tests. Don´t know how the US version compare in that respect. Raw signal from eddy probes was accessed in the cubicle in the control room in this design at the time. Olov

Thanks for all the replies and good ideas... I notice that I forgot to mention this is a RCP on a B&W RCS. Runs at 1190 RPM Allis Chalmers motor with a Byron Jackson pump, N9000 seal with no vapor stage. I had pretty much ruled out the motor due to lack of indication... We only see a slight vibe change (down!) on the lower motor probes and nothing on the upper bearing probes where the flywheel is.
I'm glad seal parameters were mentioned, as we did see a increase in seal leakage at the time of the first spike. It went from 13.5 to 16.5 gpm and has stayed elevated ever since, even though the spiking has quit. The seal has been installed since 1998. Leakage is monited via a drinking bird output and is only recorded twice a day, so I don't have actual real time seal leakage data. Can't find any changes in seal flow temperatures, pressure breakdown across stages, or any of those parameters. I've been told unconditionally by vendors that there is no way any part of the seal could exert enough forces to affect shaft vibes, but the data sure looks like something happening at the seal.
I ruled out Bently rack problems because the spiking is showing up on multiple probes including the lower motor probes which are a different rack card. Also another RCP's cards share the same power supply and comm processor.
1x phase moves with the spiking, 2x phase and amplitude are steady. 1x phase comes back to normal as amplitudes return to normal, so it is not a shaft crack or moving mass. Spectrum just shows an increase in 1x during the "spike" with a very slight rise in overall noise floor.

Thanks,
Ed

One of our RCPs suffered from changes in 1X vibration due to changes in coolant temperature. "No duh" you say. These changes, however, were very dramatic. The changes really had more to do with differences between coolant temp and seal water temp. Somehow, an annulus of water between the pump shaft and thermal sleeve is supposed to protect the upper pump bearing from high temperature. A pinhole leak at this thermal sleeve caused water temperature in this annulus to change at the location of the leak, resulting in a thermal gradient that caused a shaft bow...hence changing 1X vibration.

OK...it's a stretch to think this exact issue is affecting your pump. I offer it, however, to suggest you keep focus on the possibility that the frequency of VCT make-up may line up with the vibe spikes. Introduction of cooler seal water may cause some thermal changes? Again, would be very curious if you determine a cause.