Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #1 
I have a multistage pump with up to 8mm/s, where the main component is a nonharmonic, supersynchronous, sinusoidal vibration. I'm aiding in troubleshooting remotely.

This is a closed loop with a high-nq multistage pump, which goes through and orifice, then through a energy recovery turbine (pump-as-turbine). The pump-turbine-motor are all on the same driveshaft with no gearboxes. Turbine is in the middle so can't be uncoupled. This is all mounted on a suspended steel structure 15m off the ground (very flexible). There is a booster pump that is mounted on the ground with a separate drive.

The booster runs at a fixed 1500rpm. There is signifcant transmission of 25Hz vibration to the main pump, somehow. The main pump train runs at 2300-3000 rpm. The main pump spectrum has components at 25Hz, 1x, 2x, BPF as expected, but the main component is 108 Hz. The amplitude of this component is proportional to the pump's absorbed power. There is a slight dependence on speed and operating point (105-108 Hz), but otherwise does not scale with speed.

The 108 Hz is definitely not harmonic to either main or booster drive train. There is obviously some amplification happening somewhere, but before we account for that, we need to understand where this 108 Hz excitation is coming from. It is very sinusoidal (no haystack) with minimal 1x sidebands, so I doubt that it is broadband excitation of a natural frequency, but otherwise I have no other idea. It does disappear when shutting down the loop, so is not just external vibration exciting the structure.

The only phenomenon I can think of that meets the nonharmonic, supersynchronous, sinusoidal specification is vortex shedding? But usually, diagnosing vortex shedding is like Dr. House diagnosing Lupus. Its never Lupus.

Anything else I can consider before calling out my colleague on giving me bad info?
OLi

Sr. Member
Registered:
Posts: 1,914
Reply with quote  #2 
Touching exciting a resonance? Yes otherwise you may be in the flow induced swamp and may your deity preserve or possibly help you.
__________________
Good Vibrations since early 1950's, first patented vibrometer 1956 in the US.
http://www.vtab.se
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #3 
Quote:
Originally Posted by OLi
Touching exciting a resonance? Yes otherwise you may be in the flow induced swamp and may your deity preserve or possibly help you.


My colleague gave me some spectra. Ugly, but it shows everything. This is at 2684 rpm (44.7Hz). The zoomed spectrum is at a higher speed, but shows the sinusoidal nature of the 108Hz peak. In the case of touching,  would we not expecting haystack around the eigenfrequency?

Samp_Spec.png
Zoom of the 108 Hz peak:
  Samp_Spec_z.png 

OLi

Sr. Member
Registered:
Posts: 1,914
Reply with quote  #4 
If it is sharp enough maybe. What sensor do he use? Any knocktesting done just to check?
__________________
Good Vibrations since early 1950's, first patented vibrometer 1956 in the US.
http://www.vtab.se
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #5 
Quote:
Originally Posted by OLi
If it is sharp enough maybe. What sensor do he use? Any knocktesting done just to check?


These are PCB single axis sensors if I recall. Probably magmount. He showed me his setup a few years ago. NI system, you can probably tell from the labview graphics.

He did perform some bump tests on the pump, and the 108 Hz is about as far away as he could get from any eigenfrequency! I reckon there is a vertical eigenfrequency of the entire steel structure though. He wouldn't have seen that with his 1lb hammer, though. I also imagine there could be an acoustic eigenfrequency, so I am having him take a shot-in-the-dark of rerouting some piping. But still, eigenfrequencies aside, there still needs to be a relatively sinusoidal excitation source...
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #6 

I’m sure you would have considered, but I have to ask   - are there any rolling bearings on this beast?  (perhaps it’s a series of non-sync 1x sidebands where this particular sideband at 108hz is resonantly amplified even though it’s not the fundamental fault frequency).

RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,810
Reply with quote  #7 
I googled "super-synchronous vibration" and got a number of hits, among which was an interesting pump case: https://www.mechsol.com/case-study/elimination-of-super-synchronous-pump-vibration/

Whenever I see a distinct peak, that is not a harmonic of any forcing function that I can identify, I think "rub" because a rub, like an impact, excites all frequencies, and thus all natural frequencies.  If you can change the turning speed without changing the frequency of the peak in question, you have your confirmation.

__________________
"The trend is your friend"
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #8 
@epete - bearing FDFs and any associated sidebands around harmonic components scale with shaft speed, so I don't imagine that would be the cause. Besides, this is a brand new build, so we shouldn't have any defects in the bearing.

@rusty - Interesting. That article seems quite applicable. It says the response frequency and rotor natural frequency are different, which I don't understand, but I guess its more common to excite the rotor mode directly. No chance for a shaft vibe measurement, but I will have my colleague check the internals and see if there was a rotordynamic calculation on this pump.
Danny Harvey

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,402
Reply with quote  #9 
"It says the response frequency and rotor natural frequency are different, which I don't understand, but I guess its more common to excite the rotor mode directly."

I think I do but I'm certain that I could not explain it. I would think that the difference would be because a rub changes stiffness and isolates a portion of what it is rubbing.  I'm thinking of a bow and a finger on a string.  The position of the finger effectively changes the length and mass of the string so it increases the natural frequency.  The bow provides the exciting force but the lateral position of the bow relative to the antinode of the string would also alter the response frequency wouldn't it?  I play guitar and where I hit the strings definitely effects the response although I've never measured it.  With me playing there are a whole lot of other variables that would change things. [cool]
electricpete

Sr. Member
Registered:
Posts: 647
Reply with quote  #10 

Yeah, I missed that doesn’t track with speed part (I knew you wouldn’t overlook the bearings without a reason).

 

Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #11 
@Danny, I see what you are saying. I've never seen a 'modified' rotor mode in this way. I read that rotor modes only ring due to a rub if the incidence angle of the rotor striking the stator is quite high, therefore more of a bump than a rub. In that case, the the contact exists only for a moment, then vibrates free again. But again, I've got little practical experience with rubs.
WWST

Sr. Member
Registered:
Posts: 86
Reply with quote  #12 
Fluid flow exciting a helmholz resonance?
Walt Strong

Sr. Member
Registered:
Posts: 889
Reply with quote  #13 
More detective (and testing) work is needed for flow-induced vibration. Possible flow disturbance locations include: flow control orifice or control valve trim, piping side branches, thermowells, and pump/turbine inlet (cascade flow).

Happy Hunting!
Walt
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.