Hi everyone,

Could use some advice on this machine. This is a BB5 barrel pump in boiler feed service.... the drive train is actually a 4 machine train.... Boiler feed booster pump is coupled to the motor, the motor (25 Hz) is actually tandem coupled to a fluid drive which then couples to the main boiler feed booster pump (subject).

The fluid drive has a 4:1speed increasing gear set that is always running at 100 Hz... this peak transfers across the coupling into the main BFP (very low amplitude) and is always present.... this peak itself in the spectrum is not resonance.

The problem with this pump is the PIV vibration amplitudes (around .4 ips when I collected data)....the main BFP will usually runs around 5445 rpm (will vary speed depending on unit load), and as they approach this speed of 5445 rpm the vibration at the PIV location increases considerably (which had me thinking about a potential resonance).

*Note - this site only records DCS vibration from the DE in the PIV location... there is no sensor on the PIH location that they can see*

I have attached a PDF with all of my impact test results as well as two spectrums from the PIH location and the PIV location, and both the PIH/PIV spectrums in log scale.

The vibration at the inboard bearing housing is not highly directional.... PIV was .4 ips while the PIH was around .34 ips. The vertical direction of this bearing housing is very stiff... there is a very short/stiff 4" oil drain piping beneath it, which surprises me that this direction would have the highest vibration amplitudes.

In doing my impact tests, you can see there is a peak with a substantial phase shift right around 100 HZ. You can see this peak also in the PIH location as well.... looking even further away at the POH location, you can see a hint of the peak with a clear phase shift.... not in the POV direction though.

My initial thinking is that this resonance is very heavily damped, however it may be contributing to the amplitudes in the PIV direction. Do you have the same opinion based on the provided data? I don't believe this is the "sole problem" behind the 1x peak, but I believe it is increasing the amplitudes to a certain degree.

Operating speed is 90.75 Hz, while this peak is at around 100 Hz. That is a separation of around 10%.

Cross channel phase from the PIH to PIV locations showed a phase shift of 312 degrees, so I don't think there is much influence from imbalance.

Cross channel phase from the POH to POV location show a phase shift of 123 degrees, which makes me think there isn't a significant influence from imbalance on this end either.

There are 3 other pumps running in this building that are of the same design and speed.... I was concerned that the 100 Hz NF I think I am seeing is background noise, however given the phase shift... I am thinking it is a true NF. Coherence at this peak during the impact testing was .948. Looking at the spectrums in log scale, it does look like this resonance is present.... do you agree with this?

An additional question of mine is this.... you can clearly see the non-resonant 100 Hz peak in the spectrum... as previously mentioned this transfers across the coupling into the main BFP... this peak is very low in amplitude... if this 100 Hz NF peak I am looking at is in fact resonant, why would the 100 Hz peak that transfers from the fluid drive not be excited (ie; higher amplitude)? Technically, there is no excitation for this peak in the pump... it comes from the fluid drive. Not sure about this?

Could you please review and let me know what you think? Would appreciate any and all help.... just want to make sure I am not looking "to hard" for resonance and trying to make something out of nothing.





Original Post


Two cents worth,

Can you provide a multiple point plot using as many horz, vert, axial readings on the machine train that you can plot? Example shown below HVA on 6 bearings on a motor/screw compressor. This helps to understand which part of the machine train has highest vibes and which axis is affected the most. Got any vibe/maintenance history of any of the machines?

Your bump test data, in my opinion, the PIV, no POV, bump does not show a 180ish phase shift, shows 80ish phase shift instead. The PIH and POH show 180 and 120 respectively. I would say this is a natural frequency. Bearing housing/pump shaft/base/piping. Search of BB5 barrel pumps pulls up a whole bunch of different designs, need some more info/pics of your specific installation.

When you say it is not highly directional, this leads me to believe the bearing housing is orbiting in a circular pattern, rather than a 3-1 to ratio or moare from H to V.

So there is a sensor, has the vibration gone up recently?

What about the other 3 machines, do they exhibit the same vibration at the same machine train component at the same speed?

So on the problem machine, you maybe seeing vibes from other 3 machines, if you can change speed on one of the 3 machines while monitoring problem machine in a steady state, does the vibration on problem machine change?

Your posted vibration data PIH 0-400, at the base of the 100 hz peak you can see a small hay stack, when you look at the other peaks they do not show that, this hay stack is one indication of resonance at that specific frequency.

Not sure about your description of how vibration is traveling across the coupling? Could be base or piping transmitting the data. Have you done phase from the base to foot to bearing radial and axial? I would anticipate .009 in/sec or less on base readings, either overall or at any frequency. 

Any process data changing at this speed, like a valve, damaged check valve, temps, etc.



Photos (1)

Hi Dave,

Thanks for the reply on this and for providing your thoughts.

In response to your questions:

1. I use a VX2 analyzer with Omnitrend analysis software.... to my knowledge.... I am unable to plot a similar plot to that one which you provided the example of.

   1a. I have very limited historical data....  the one thing that I do have which continues to suggest resonant "contribution" is a historical plot that shows speed vs 1x amplitude over time.... from 5300 to about 5450 rpm, the DE vertical positions amplitude increases quite a bit, ie; 5250 rpm (1x) at .188 ips, and 5300 rpm (1x) jumps up to around .295 ips and stays elevated until about 5500 rpm, at which time the 1x decreases back down to around .188 ips.

2. I would agree with your comment that the DE bearing housing may be in an orbiting type motion. The thing that I can't wrap my head around is why the PIV amplitude (.39 ips) is slightly higher than the PIH amplitude (.35 ips) given the resonance appears to be horizontal? Maybe this is purely due to the mode shape? Horizontal is definitely the least stiff direction on this bearing housing.

3. The "sensor" they installed which ties into their DCS/PIE system was only recently installed.... this is how they came to find out there was even an issue here.... historically they had not been collecting data on the DE bearing housing. The vibration is consistent.... what I mean by that.... it "jumps" up in amplitude when they get to around 5300 rpm as mentioned above.

4. Regarding the vibration that transmits across the coupling:

The fluid drive has a 4:1 gear set, and then the scoop tube style fluid drive stage. The 4:1 gear set is always running at 100 Hz, and they then vary the scoop tube position for output speed of the main BFP..... that 100 Hz peak is always present in the pump vibration.... it comes across the coupling. If you look at page 4 of my attachment on the PIH 0-400 HZ spectrum, you can clearly see the 1x around 90 HZ and then the small 100 Hz peak to the right. That is a non-resonant peak that comes from the fluid drive 4:1 gear set.

5. I have taken phase from the foundation to the feet on the pump.... there does appear to be some minor degree of looseness present, and the amplitudes are a little higher than I had anticipated.....any where from .02 on the concrete/grout to .09 ips on the pump feet themselves....I have always struggled to understand when to make a call on foundation looseness.... I know you look for the phase shift... but not sure at what amplitude to actually say this is an issue. Does that make sense? What has been your experiences with foundation looseness and when to say "this is an issue"?

6. Regarding the performance of this pump, I have trends that show pretty consistent conditions at this load rate.

Hope this answers some of your questions. Really appreciate the help.




It has been my experience that values over .02 ips is out of tolerance, either base or machine feet. If you "map" the base around the feet, you will most likely find it is localized to 1 or 2 feet. So your overall/1x amplitudes on the base should not exceed .02, as I stated before I like to see .009 ips, you may find these values as you move your sensor in vertical direction on the base. 

I agree there is a natural frequency, question is what. Is it the base, pump foot, pump shaft, pump impeller, bearing housing, worn coupling.

What about the other 3 machines, do they too exhibit the same vibration amplitudes at same bearing locations?



Add Reply