1x vibration on centrif compressor

[Stevemac]

I have a 2 centrif compressors operating at 8500rpm, driven by solar gas turbine.
Experiencing 1x vibration with no significant sidebands at coupling end of 2nd driven compressor.
Coupling spacer shaft 36" Kopflex.
Vibration does seem to increase with speed, but what other possible causes besides balance.
Keep it simple, vibration is not my field

Steve

 

[moygr1]

Allignment? Joseph Moylan
moygr1@home.com

 

[sms]

High 1X on a centrifugal compressor can be a symptom of many things, balance, alignment, resonance, shaft crack, rub, excessive clearance,just to name a few. You really need to have someone do a complete vibration analysis to get to the cause.

 

[electricpete]

For real-world machinery vibration questions, try the vibration/alignment/balancing forum at:

http://www.reliability-magazine.com/ubb2000/cgi/Ultimate.cgi?action=intro

In my humble opinion, they have many more hands-on practicing professionals in that forum. Although perhaps not as many theoretically-oriented folks as you'll find on this forum.

 

[Shaker]

I would check the simple stuff first. 1. balance the machine. 2. Check the alignment.
Still not fixed check for resonance.

 

[brin]

If you do go for shut-down, disassemble and inspect the couplings to make sure nothing has moved that could affect balance or shaft alignment (eg: loose retainers, binding,etc.).

If the unit is still running, what is the direction of strongest vibration (axial/horizontal/vertical)? What do you have available for instrumentation?

 

[Hatch]

Was this a new installation?
Did the vibration problem develop suddenly or over time?

All the possible causes are covered in the other suggestions. A full diagnostic vibration measurement is required.

However, if this was a new installation, and the equipment was sitting at site for a long time prior to installation and startup, I would check rotor bow, which would appear at 1X and 2X.

 

[blassch]

Steve,

in some cases i have seen certain types of elastic element couplings produce 1x cyclic forces.

Is it a new installation ? What kind of coupling ?

Christian

 

Hi,
Can you tell me
1.did you captured bode plot for start/coast down?
2.what is the value at other measuring planes of amplitudes and relative phase angles? ( Phase angle is important )
3.Check Looseness because sometime problems like small unbalance,etc is aggravated(increased) by looseness of foundation bolts at 1X freq.
4.did you consult Manufacturer, Try to get reference vector or fault frequencies.

I am waiting your reply.
Thanks.
K.Muhammad

 

[Franko]

When vibration is high only on a coupling end, the coupling is a prime suspect. There could be damage only on one end of the coupling. If coupling or the fit area are not damaged, coupling unbalance could be checked by vector changes with 180 degree rotation of the hub if it is keyless or has two keys (instead, a spare coupling should be used if available and have Koflex perform the balance check).
It is suspicious that vibration does not change with speed. If vibration data uses shaft eddy current probes, check instrumentation and make sure that bearing housing readings have also increased. As previous posts indicated, shutdown data with Bode plots (amplitude and phase angles versus speed) will show the true change in "1X" vibration. Eddy current probes can have false readings which can be seen at slow roll speed. See

http://www.bently.com/articles/apnotes/an002.asp.

(Bently Nevada also offers service and problem analysis as their other publications show). However, if a rotor becomes bowed, mechanical "runout" is a real problem - the probes are acting as dial indicators near the bearings. Slow-roll 1x amplitude and phase data with the rotor in its initial straight condition is required to tell if there really could be a bow. Also, a bowed rotor will cause high excitation at the rotor/bearing first critical. It usually isn't the cause of high vibration only on one end - well above the first critical - unless there was a rub and the rotor is "kinked". Bowed and "kinked" rotors are why runout subtraction instruments never took off years ago - there were too many misapplications.

 

[ronfrend]

I had a lot of experience on Solar GT/compressor sets and you can get lots of problems.

You have identified that the prime vibration is at 1x. You now need to take orbits of the XY sensors. If you haven't done this before all you need is a scope with X-Y coupling capability. Connect to your Dymac or Bentely Nevada outputs. If the plot is circular then you have imbalance. If the plot is elliptical (or even figure of 8 shaped) then you have misalignment. If you have a deformation of the plot either as an internal loop or a bite out of the plot then you have either a rub or a "rotating stall". If it is a rotating stall then you will also see another frequency about 0.6x run speed.

Another prime cause of high 1x vibration on Solar centrif compressors is wear in the bearings. As the pads and/or pins wear down then you have additional clearance allowing extra movement and as you are using displacment probes, you will see more movement of the rotor shaft which will be interpreted as imbalance.

I was caught out once by a scratch on a shaft which gave a high 1x. However if this is the case you will see it on the XY plot as a deviation on a small circle.

If you have to balance the rotors make sure that the technicians who reassemble it in the field are scrupulously clean - only a little dirt can seriously affect the balance.

Hope this helps

ron.frend@predicon.net