Sign up Calendar Latest Topics Donate
 
 
 


 
Poll Results
 
 Why is this data impossible?
 The "TRUE PEAK VEL" values must be more than the "RMS VEL" values 9 100%
 The Horizontal vibration must be higher than the Vertical vibration 0 0%
 The radial vibration (H or V) must always be higher at the DE than the NDE 0 0%
 The vibration at the DUTYPOINT (assumed BEP) must be lower than at MIN FLOW 0 0%
Total votes: 9. This poll has been closed.


Reply
  Author   Comment  
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 471
Reply with quote  #1 
Here is one of my favourite questions from my training material.

Attached is vibration data for a simple centrifugal pump during a factory acceptance test.

Just by looking at these values for a few seconds, you can figure out that these values are impossible.

How so?

YammerVibeLevels.png 

Barry

Avatar / Picture

Member / Supporter
Registered:
Posts: 112
Reply with quote  #2 
IMO BECAUSE it is made up to bolster you point of view.
__________________
Barry
John from PA

Sr. Member
Registered:
Posts: 810
Reply with quote  #3 
Are these values all obtained with the same instrument or two different instruments?
fburgos

Sr. Member
Registered:
Posts: 443
Reply with quote  #4 
I wonder whats the setup, is the sampling rate or fmax different for both?
RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,566
Reply with quote  #5 
Curran, could you not have data with broadband, low amplitude “noise” that would give a higher RMS value while giving a lower True Peak value?
__________________
"The trend is your friend"
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 471
Reply with quote  #6 
I didn't want to put in any caveats on the RMS vs. Peak measurements because that would have given the answer away.

This is real data, but it was in Mumbai, so who knows what the hell they were doing down there 10 years ago. I think they interpretted 'true peak' that they needed to fill in the highest peak in the spectrum. We do use 'peak' to describe a lot of things in our business.

Rusty: in the extreme case, a square WF, the true peak and RMS are equal. With a sinus, you all know the true peak is sqrt(2) higher, and with broadband noise (say, white noise), the ratio is higher, but because its stochastic, it depends on the length of the sample. Any combination of these will still result in a higher true peak than RMS.

The question assumes the machine is completely steady state, the sensor is the same and analyser has the same settings. Lets just say its coming from the same WF.

fburgos

Sr. Member
Registered:
Posts: 443
Reply with quote  #7 
could they be mixed up, all rms values are roughly 2 or 3 times higher than "true peak"
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 471
Reply with quote  #8 
Quote:
Originally Posted by fburgos
could they be mixed up, all rms values are roughly 2 or 3 times higher than "true peak"


We don't make pumps that can vibration as low as 0.28 mm/s RMS. [rolleyes]
fburgos

Sr. Member
Registered:
Posts: 443
Reply with quote  #9 
could be one masive slow speed pump, i've measured some horizontal 20MW hidro generators with overall as low as that
electricpete

Sr. Member
Registered:
Posts: 517
Reply with quote  #10 
I vote for the first one. 

Assume that RMS and true peak are determined from the N samples.   (Other signal processing options would need separate consideration. )
Let's say xmax is the true peak magnitude, xi are the sample magnitudes for i = 1..N
By definition of xmax:  xi<=xmax for all i  
Exclude signals whose absolute magnitude never changes (example DC and square wave), therefore:  xi<xmax for at least some i  

RMS = sqrt(sum(xi^2)/N) < sqrt(sum(xmax^2)/N)   = sqrt(N*xmax^2/N) = sqrt(xmax^2)=xmax
RMS < xmax
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.