Piping Resonance 5

[teowl]

Hi all,
I have a long standing vibration problem which I hope to share and hopefully can get some good feedbacks on how I can resolve the issue.
We have a centrifugal pump system. During operation, the pipings are vibrating very badly, as much as 0.8 in/s. FFT Spectrum showed a clean 4 x rpm peak, which happens to be vane pass frequency. Immediately we suspected the natural frequency of the piping to be also at 4 x rpm (which is 200 Hz), and that resulted in resonance of the pipings.
I attempted to do bump test on the piping with a copper mallet, but was unable to see any peak at 200 Hz. Is bump test the right approach here to determine the fn of the complex (many elbows) piping? Can I get the pipes to resonant at 200 Hz with a copper mallet?

Appreciated any comments

 

[sms]

You should be able to get a 200 Hz resonance with your mallet. It may not be pipe resonance, it may well be that the vane pass vibration is high due to recirculation in the pump. Might start by figuring out where you are relative to BEP.





"Why don't you knock it off with them negative waves? Why don't you dig how beautiful it is out here? Why don't you say something righteous and hopeful for a change?" Oddball, "Kelly's Heros" 1970

Please see FAQ731-376 for tips on how to make the best use of the Eng-Tips Forums.

 

[GregLocock]

Switch the Hanning or Hamming filter off when doing an impact test.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[Tmoose]

I'd expect a large non-metallic mallet or A dead blow sledge hammer thumping at least several diameters away from the accelerometer, or cutting a preloaded wire at a point of high operational vibration may provide an impulse more effective at exciting pipe frequencies. I like to see the pulse and decaying ring down in the captured time wave form. If worse comes to worst the big frequency can be measured or confirmed off the waveform.

What is the impeller diameter? If it is "full sized" then trimming the OD (even better with a bevel) or the cutwater can soften the blade pass pulses quite a bit. Like others said, legal flow conditions should be relatively free of blade pass pulses.

 

[ApC2Kp]

teowl,

The 200 Hz observed vibration is probably a higher order harmonic of a much lower natural frequency of the piping. Your hammer test will ring the first and low modes before much energy gets to the 200 Hz mode shape. Your pump forces the piping to vibrate at the vane passing frequency, and is too high to excite the fundamental frequency. Typically any piping with visually observable vibration has only a few supports to carry just the weight of piping, and does not have restraints to limit the lateral movements of the piping. Such 'flexible' piping systems could have natural frequency less than 2 or 3 Hz, and would have large response to wind gust, waterhammer, or other occasional loads such as seismic event. An adequately restrained piping system might have a fundamental frequency of at least 10 to 15 Hz.

If the pulsation from pump vane passing frequency cannot be modulated with a pulsation dampener (Blacoh? how much attenuation at higher frequency of 200 Hz), then the piping supports should be modified by added restraints. A hot piping system will need to be analyzed for thermal movements to provide adequate gaps at each restraint point to allow expansion, yet to close the gap at operating temperature for a rigid effective restraint. The restraint structures will need to be rigid - a threaded rod or cable support does not provide much lateral restraint stiffness.

A rule for minimum of one limit stop with two guide restraints per run of piping would serve to restrain the piping for seismic events, and would also serve for a resonant condition piping system. Harmonic analysis of the piping system could provide a more exact design for pipiing restraints, but the general rule of one limit stop with two guide restraints could be sufficient for a smaller diameter piping system NPS-3 or less. This method for restraint design could result in triple the number of supports on the piping. Each support allows the vibration energy to be dispersed so that there less excitation in the piping system after a short distance (and several restraints) from the pump.

 

[Zimann]

teowl,
I remember a similar issue I had, but first a comment.
"Long Standing" issue - does this mean that it was there from original installation, or has it developed/worsened over time. I will assume it was there from the beginning, but if not, then an inspection/repair report (of the vane tips, wear rings, etc) may reveal good info.
I had a between bearing, single stage, top suction, top discharge, eccentric reducers on suction and discharge, 5 closed vane, 1800 rpm, low suction, high discharge, high flow unit charge pump from a company in Europe that starts with an S that had 5X vibe problems from original installation. Actually, the company I was with had installed 2 identical pumps into 4 identical systems, and all 8 had similar problems.
We checked out the pumping system, the foundation, and the pump construction.
Ultimately, we reviewed all info with the vendor and they agreed to provide new impellers (but first, we attempted modified "B" gap, modified tip profile, and added mass to the bearing pedestals - none of which satisfactorily eliminated the problem) While they had quoted a design that should be sufficient for the service, it was too susceptible to vane/cutwater interaction.
You could discuss your specifics with the vendor.

 

[Tmoose]

I think there is usually an advantage of providing excitation at a point confirmed as having high operational vibration, or even a point of max vibration (antinode). If I inadvertently try to excite a node then the mode and frequency of interest may hide out under the dock as big fish often do.

It may be interesting to map a few points and directions of high vibration on the piping starting close to the pump. I struggled with a few systems where pressure pulses were able to make the elbows flex and thus give the straight piping runs a lateral shake to start and drive piping resonance.

 

[teowl]

Dear all, Thank you for all your valuable input.

Zimann, I think my problem was there since day 1, but that was many years ago before I was here. Recently, the issue seemed to deteriorate. So what was the final solution the vendor gave to you? Did you manage to find if it was the piping resonance that caused all the problems? In my case, the further I operate from the pump BEP, the worst the vibration become.

Anyone has any experience in changing the number of vanes in the pump impeller to resolve similar issue?

 

[Tmoose]

"Anyone has any experience in changing the number of vanes in the pump impeller to resolve similar issue?"

Pipe resonance >>> "excessive vibration" , no.

But Yes, in a medium rise office building when blade pass frequency was reappearing here and there and causing noise complaints thoughout the building. The real issue was poor pipe mounting in some of the vertical chases, but an impeller change was financially more attractive.

 

[Tmoose]

Depending how much flow you really need a VFD might be used to change the rpm, and thus blade pass frequency dirtectly, and maybe even keep the pump in a better flow regime

 

[ronfrend]

Just to add a little more to the subject.

I have also found (more than once) that the vibration is often linked to an acoustic vibration in the pipe. Using c=fl
where c = speed of sound (1500 m/s in water)
f = frequency (200 Hz )
l = wavelength

you can quickly work out that if you have a length of pipe of about 7.5m that will give a fundamental for an open/open pipe. Be careful though - the vibration is often at an "overtone" which is NOT a direct multiple of the fundamental.

Good Luck

Ron

Ron Frend

http://www.predicon.net

 

[HookedOnHalo]

The 200 Hz mechancal natural frequency could be a shell mode of the pipe, hard to excite with a hammer unless you do the test right with the right xducer.

What size is the pipe? Wall thickness?

 

[Zimann]

Sorry about the delay.
Yes. 1 of the 8 pumps now has a new impeller - they want us to pay for the other 7. The new impeller is designed for a BEP at a slightly lower flow - but still sufficient for operational requirements. The fix was not perfect, but a significant improvement.
My recommendation is that you review the operation requirements for the pump and discuss those details with the vendor.
I have usually found it to be cheaper to modify the internals of a pump than to re-design the piping. (modifying pump internals includes: better selected impeller, modified cutwater, modified rings, etc) This, of course, is only if operational requirements can support it.
Tmoose mentioned a VFD - that could be great for you in a number of ways.

 

[Tmoose]

Was the new impeller different (smaller) diameter, different number of vanes, or both??

 

[Zimann]

1) Lower capacity impeller (smaller diameter). Sized for better performance where the operation spends the majority of the time while compromising slightly on the high end of the previously specified flow range. (Operations still operates at the same flows that they used to.)
2) Changed to a staggered vane design, so there is effectively 10 discharge vane tips. (Didn't previously mention in thread that it is a double suction)
3) Ground the casing volute lips.

Basically, the changes helped reduce the forcing functions of the resonance. Still has vibes, but now at an acceptible level.

 

[teowl]

Hi Zimann,

Just about right time to see your reply. I have spoken to the vendor, they too suggest some of the modifications done on your pumps, such as changing the vane angle to soften the pulsation. And also staggered the vanes so it become 8 vanes to produce pulsation of higher freq but much lower amplitudes. Grinding the volute lip is also an option. Not looking at trimming the impeller because of head required.

One question...was the vibration bad on your pump bearings? In my case, vibration is not so bad at the bearing and casing, but terrible on the suction and discharge piping.

 

[Zimann]

It was high everywhere - and loud.
Probably had the wrong design (probably should have had a multi-stage) but the Project Group who installed it opted for the cheaper version. This "fix" was also the cheaper of the possible solutions. You get what you pay for (hopefully) and our fix was acceptible, but not perfect.

 

[teowl]

Hi Zimann,

Looks like we there are a lot of similarities to our problems. Your information gave me more confidence to what the vendor proposed. Like you say...may not be perfect, but it will be good enough to achieve significant improvement. Thanks

 

[Rob45]

Why not try a resonator, e.g., a piece of pipe 147-5/8" (3.75 m) long placed about double that length from the pump?
This should provide you with a quarter-wave tuner that will cancel the 200 Hz pulsation.

Alternatively, you can design a Helmholtz tuner that would take up considerably less speace.

 

[PulseMaster]

You likely have an acoustic resonance, as vane pass pulsation is usually low amplitude. Cross wall frequencies of big pipes, or fairly short stubs would fall into this range. If there is a mechanical frequency involved, it could also be a breathing (shell) mode of the pipe. In this case, changes in pipe size, or changes to the pump case could help you.