Double Suction Centrifugal Pump Vibration

[dshiel]

I work in a Paper Mill and one of our Fan Pumps Has developed high vibration readings. The Pump is a Goulds Double Suction Centrifugal Pump 3400 series. It has a Paraflex Coupling. The Pump runs at 1770 RPM. In the past month we have developed high vibration in the pump. Most of the vibration shows up as 1x. The running element was changed out twice and the pump still has the high 1x vibration with little difference noted between running elements. Readings are as follows in veloscity: Motor- MAH.0625ips, MBH-.0941ips, Pump- PCH-1.22ips 6.6ge PCV- .53ips 6.6ge, PDH- .49ips 9.4ge, PDV- .16ips 6.4ge, PDA-.74ips 8.0ge. The FFT's show a large 1x spike and a small 2x spike. Phase readings showed 171'. Recently, last two days, the Paraflex Coupling Insert failed. Does anyone have any ideas of what is causing this vibration to suddenly occur and continue even after changing the running element????

 

[electricpete]

what is a running element?

If I ignore the high gsE readings, I would tend to think of misalignment (possibly pipe strain, thermal growth, soft foot, etc). Things that point to that:

- *coupling insert damage
- 180 degrees phase shift across coupling (What orientation was that?)
- 1x plus some 2x.
- Axial vibration.
Normally I would expect to see some evidence of misalignment on the motor. Did you get vertical and axial readings at the motor?

I'm not sure about the gsE readings. Note that the location of the highest gsE reading does not fall at the lcoation of the highest velocity reading (3H).

Perhaps those gsE levels are normal flow noise for this pump? Do you have any other similar pumps? You might take a quick check for harmonics of bearing fault frequencies. Also a timewavefrom looking for impacts or other things to help determine whether gsE levels are normal or indication of a problem.

 

[micjk]

dshiel,
I would have to agree with pete. How was alignment done.!80 deg phase shift across the coupling axially normaly signifies misalignment. Have bearings also been replaced, you may have a "cocked" bearing. do you have a multitude of 1X harmonics. I assume the "running elements" are impellers of some sort. Are they possibly skewed. I this is an overhung type situation, it would give you a high 1X and possible 2X with other variables Roy Gariepy
Maintenance and Reliability Dept.
Bayer Corporation Dorlastan Fibers Div.
Goose Creek, South Carolina USA

 

[dshiel]

The running element is the shaft, half coupling and bearings. The pump bowl is split so you can remove the top half (Bonnet) and replace the running element. The bearings set into a saddle arrangement with a boss so I don't believe they are cocked. Motor readings in the Vert.= .04ips 3.0ge, axial= .09ips 3.3ge Vibration reading taken directly from the bowl are: Horz.=.49ips 4.1ge and vert.=.34ips 20.6ge Bearing temps are 140'F to 150' F they are lubed by grease.

 

[vanstoja]

It is possible that that you have a piping system resonance at or near the 29.5 Hz running speed of the pump which has perhaps shifted onto your pump running speed due degradation of piping supports? 1800 and 3600 synchronous RPM pump piping systems typically have flexural frequencies close to pump running speeds. Does your spectrum show you the broadband level of the signal at the 1X frequency which might show the discrete 1X peak popping out of a wider bandwidth peak, ie a piping structural or hydroacoustic resonance? For many years , I have been disconcerted by seeing published frequency-amplitude spectra of machinery that do not show the signal broadband level but only show peaks emerging from a straight-line that is the bottom of the amplitude range scale. Our signal analysis practice is to NEVER let any single peak level control the analyzer dynamic bandwidth in such a way as to cut off the broadband signal in the frequency range of concern. Such peaks should be cursored for amplitude and frequency and then deliberately cut off at their top to enable the broadband
to be viewed.

 

[VibEngr]

dshiel,

Have your operating conditions changed in the last month?

Is the motor variable speed?

Are you running at a different point on the pump curve?

Has the consistency of the process fluid changed?

When you check the motor current, is it pulling more amps than before?

 

[mdimond44]

After changing the element twice, it would almost be safe to assume that the shaft/impellers weren't the problem. Notice I said almost. I would be looking at the alignment based on the high vibrations and the subsequent coupling failure. The only thing that I could imagine might have changed to move the pump would have been piping. It might be worth your while to walk down the piping and check all supports, hangers, etc.

 

[dshiel]

Thanks for all Your suggestions, on December 10th we installed a new pump to find that the pipe supports on the discharge side of the pump were in need of repair and had allowed the pipe to drop. Couldn't get the Maintenance manager to check alignment prior to this. Next time hopefully he will listen. I am just use to seeing misalignment as running speed with several harmonics. Again thanks to all for your suggestions.