Very high sub-synchronous vibration on vertical pump 1

[ME1289]

Hey Guys,

Figured I'd give this forum a shot since this particular problem has us really baffled.

We have three 3600 RPM / 200 HP vertical 8 stage pumps pumping out of two LPG spheres. At any time two of the three pumps are running. One pump moves a sphere to the unit and one pump recirculates the other sphere. Each pump has their own suction and discharge piping and the in service pumps are periodically swapped back in forth based on demand.

Since commissioning, we've seen intermittent vibration show up on all three pumps. The vibration is at a frequency of 705 CPM and ranges from 0.5 to 0.8 IPS RMS at the pump discharge head. The vibration is only in the direction parallel to the suction and discharge piping.

When a pump is started up, one of two things happen. It either doesn't vibrate at startup and will continue to run indefinitely without vibration until we shut it down. It runs fine across all flow rates, suction heads, and pump speeds (these are on VFDs).

OR, at startup, the pump will begin to vibrate almost immediately and will continue to vibrate until we shut it down. The vibration persists across all flow rates, suctions heads, and pumps speeds.
I had the suction strainers checked and no material was found.

I had a bump check done and the first natural frequency is close to the vibration frequency we're seeing. The bump test told us the first natural frequency is 750 CPM.
I believe we are exciting this natural frequency but I cannot figure out why. Even though the pumps are on VFDs, we run them at 95% output all the time. The only time we ever change speed if for troubleshooting. I can't find anything in the system that would excite a natural frequency around 750 CPM. I can't find much info online about 0.20x vibrations

There has to be something process or flow related because sometimes we'll shut a vibrating pump down for a few days, and when we restart it, it runs great.

Anyone have any troubleshooting suggestions? I've reached out to the vendor but they haven't been helpful.
Attached is a vibration spectrum I pulled when one of the pumps was vibrating.

Thanks

 

[JJPellin]

Vertical shafts in straight cylindrical bushings are inherently unstable. This instability would normally be seen as a whirl at a frequency just below 1/2 running speed. In your case, the fact that there is a structural natural frequency at about 20 percent of running speed, may be causing this whirl to lock in at that lower frequency. We have very few vertical pumps with VFD's. VFD's can tend to excite odd frequencies. But, the fact that you say the vibration continues even if you change the speed tends to discount the VFD as the source.

I would add structural stiffness to the pump head in the direction of the vibration to drive this natural frequency higher. I would modify the line shaft bushings to tighten the clearance by going to non-metallic materials. I would consider adding vertical grooves to the line shaft bushings to break up the whirl. I would consider converting to a tri-lobe triangular bushing design (reference a design by Malcom Leader, Applied Machinery Dynamics Co.).

Johnny Pellin

 

[electricpete]

lps to Johnny, who knows a lot more about pumps than I ever will.

I'll add just one thing the textbooks say about oil whirl/whip (which might be different than process fluid whirl/whip... I don't know...I'm open to comments):

Typically the vibration frequency increases with speed (oil whirl) as a constant fraction of speed up to a certain threshhold which is a resonant frequency. Then above that speed, the vibration remains at a constant frequency (oil whip). If you then decreased speed the vibration would initially stay at constant frequency until you went below that same threshhold speed where the vib frequency would again start decreasing as a constant fraction of speed.




=====================================
(2B)+(2B)' ?

 

[JJPellin]

Pete is correct. I have never been a purist in distinguishing between whirl and whip.

Johnny Pellin

 

[ME1289]

Thanks for the replies.

Just to add a few more comments, the bowl and impeller wear rings are Vespel and SS. The vertical line shaft bushings are carbon. The installed clearances are 0.010".

I finally got the vendor to acknowledge we have a problem. They are designing a spring plate to sit between the pump and motor to lower the natural frequency. Does this make sense to you guys?

Thanks

 

[JJPellin]

I have never heard of a spring plate and do not know how is lowers the natural frequency. You can simulate the effect by adding mass. Bolt plates to the sides of the motor flange using the motor attachment bolts. Stack sand bags on the plates. This will drop the natural frequency and you can prove that it will work.

Johnny Pellin

 

[JJPellin]

Those bushings clearances seem too loose. That increases the chance of whirl or whip. What is the shaft diameter at the bushings? I would expect a clearance for carbon bushings closer to 0.005" depending the shaft size.

Johnny Pellin

 

[ME1289]

The vendor is working on the spring plate design at the moment. The only reference I can find to it online is here about half way down the page:

Link

The shaft diameter is 1.686" at the bushings.

I'll wait and see what the vendor comes up with in terms of this spring plate design, it will certainly be easier to install than new line shaft bushings.
I'll keep the bushing option in my back pocket if the external modifications don't work. It sounds like we are definitely exciting our natural frequency from oil whip in the line shaft bearings.

Thanks for your reply

 

[JJPellin]

My spec for those bushings would be 0.005" to 0.007" at that diameter with carbon. Your natural frequency is already at 20% of running speed. I am not sure that lowering it more will help.

Johnny Pellin

 

[Strong]

I have seen the spring plates installed on three vertical pumps. The first design require modification leading to removal and machining. The spring plates did not perform well, so they were all removed and the motor support and pump column were stiffened as the final fix. Be very careful about use of spring plates if the motor thrust bearing also support the pump rotor! If all modifications are the responsibility of the OEM, then stand back and watch, but if down-time is an issue, then get involved!

I did detune LNG tank pump (only one of two with high vibration). The test/demonstration weight was diver's belt lead weights strapped about motor, and the final weights were bolted to supports attached near the motor base. The vibration mode had greater motion at the motor base rather than top of motor.

Walt

 

[ME1289]

Thanks for the info, Walt.

How much weight did you ultimately add?

 

[electricpete]

When a pump is started up, one of two things happen. It either doesn't vibrate at startup and will continue to run indefinitely without vibration until we shut it down. It runs fine across all flow rates, suction heads, and pump speeds

It brings to mind a situation where foreign material is held in position while running but has an opportunity to reposition during stop/start cycle. when it ends up in a position within the flow stream, maybe it creates turbulence which gives broadband vibration exciting the resonance. Such condition might possibly be evidenced by decrease in performance. Is the performance (taking into account speed and lineup) reduced when the pump is vibrating?

How long ago was the last overhaul? I assume clearances were checked at that time. Did the condition exist prior to last overhaul? Did the condition exist immediately after last overhaul or appeared slowly?

What was the speed of the machine when that spectrum was taken? Do the other peaks in the spectrum follow any pattern? (maybe they are multiples of 752cpm?)

=====================================
(2B)+(2B)' ?

 

[ME1289]

These pumps have only been in service about 2 years so they haven't been removed for an overhaul yet. I've cleaned 2 out of the 3 strainers and didn't find any significant debris.

Thanks for the reply!

 

[Strong]

705 cpm is quite a low frequency for structural vibration mode of the motor on support pedestal, but perhaps a vibration mode for the entire pump column. The LNG pump I worked on had about a 40-ft long pump column, but the natural frequency near 3600 cpm (60 Hz) was dominated by motor motion. The detuning weight was about 50 lbs. (from faint memory). I would not even consider this for your machine without more information!

Do you have any other indications coincidental to the high vibration at 705 cpm?
Have you looked at vane pass frequency for increase in amplitude, and modulation sidebands at 705 cpm or at 1xSS?
Have you made high frequency acceleration measurements to possibly detect Cavitation or rotor impacting?
I assume the pump suction is open inside the tank, so are there baffles to prevent swirling and vortices that could cause pulsation?
Is there any discharge pipe vibrations or dynamic pressure present? Any indication of gas/vapor in discharge pipe and system? Is there any possible feedback from a reciprocating gas compressor?

Regarding the spring plate: has the vendor convinced you that the motor vibration mode is dominant?

Walt