Interpretation of vibration spectrum

[gwkwong]

My Client has a recip compressor unit that is experiencing vibration issue. The overall vibration magnitude is about 0.5 ips at 1000 rpm at the distance piece support I beam. However, the manufacturer review and conclude that most of the vibration is beyond 2X operating frequency, which means it could relate to pulsation issue. And the vibration at 1X and 2X is low, which is related to mechanical vibration. I wonder if this is a right interpretation.

 

[StrykerTECH]

0.5 ips is a universal measurement of vibration. I would focus more on the frequency and accelation factor of the vibration. Then you could make a determination of where it is coming from. IPS gives me the velocity, 0.5 IPS at 10 hz is a whole lot different than 0.5 IPS at 200 hz.
That being said, the manufacturer's conclusion is correct, but how can he be so sure that the frequency is beyond 2x operating frequency? do you have that information. 1000rpm is about 17hz, so what is the frequency vibration that you are seeing?

StrykerTECH Engineering Staff

http://www.stryker-tech.com/

 

[Hatch]

Not much information to work with so I have questions.
Is the vibration problem on the compressor or is it the support beam?

What is the overall vibration at the compressor? Higher or lower than the support beam?

I agree with StrykerTech that frequency info is required.

Taking a shot at this with the limited info available, I disagree with the manufacturer. If this were a pure rotating machine, then I would agree with them. (You would not expect anything other than pulsation to be higher than 2X running speed). However its a reciprocating machine with more than one cylinder so its going to produce multiples of the rotating speed which may 1X, 2X, 3X, 4X etc of the runnig speed. Those high multiples my just be the source of the support beam excitation.

Or.....The support beam may just be vibrating at its natural frequency wich is excited by almost anything from the compressor since its probably producing broadband vibration frequencies from the pulsation and operation.

More info is needed.

 

[gwkwong]

The unit is a recip compressor. I have attached the spectra for the comp/cylinder and the I beam from the vibration monitoring. The spectra for the I-beam shows vibrations at high order of operating frequencies which I think is related to pulsation. Any idea will help.

 

[gwkwong]

here is the spectra for the comp and cylinder

 

[tensor47]

These are nearly identical spectra that I get long term on a 2 cylinder horizontally opposed recip. Here is a section from my monthly report. 1 ips rms = 25.4 mm/s rms.

Machine internals are ok.
The HP (v8V), cylinder head vert vib slowly varying around 13 mm/s. LP head vib 3.5 mm/s never changes (v7V).
Recent offshore work suggests this is a long term resonance problem, corresponding closely with 9x shaft speed, with the mass of the cylinder & head bouncing on the stiffness of the concrete column. The levels are 3 to 4 times the LP head, and cause the foundation bolts to pull loose every few years. The LP head and foundation is bigger than the HP head so modal characteristics will be different.
Ideally, modal analysis and an FEA model are needed to tune it to 8.5 or 9.5x and the level will drop. It’s not essential as it has been doing this since 4/01, but not previous to that.

There is another possibility which is an acoustic resonance of the inlet or outlet bottle and pipe work which can be reduced with orifice plates in the nozzles.

Both acoustic AND structural resonance could exist together.

Pressure measurements Vs crankshaft angle in the pipes and bottles should not be difficult, to prove the fault is accoustic, which appears to have a simple solution, and can avoid a final failure of the concrete foundation.

 

[StrykerTECH]

Your data seems to point to a pulsation issue. I am going to concur with the manufacturers analysis.

StrykerTECH Engineering Staff
Milwaukee, WI

http://www.stryker-tech.com/

 

[StrykerTECH]

Dug a little deeper for you. Tensor47 is on to something.
I think you may have an acoustic problem.

Check out the link. Your wave data looks just like the acoustic description.

StrykerTECH Engineering Staff
Milwaukee, WI

http://www.stryker-tech.com/

 

[gwkwong]

thanks very much, I wonder if there is a good book discussing vibration spectrum for recip compressor.