Vibration of Centrifugal Compressor

[bk8548]

Hi, I'm bae working for the gas production field.

We are now undergoing the field performance test of new compressor package for the natural gas.
The centrufugal type compressor has a capacity of 72,000 kg/hr, the inlet pressure of 45 barg, and the compression upto 106 barg. It is driven by a 5MW gas turbine.

We are now in the stage of the flowrate of 25,000~50,000kg/hr which is the flowrate below a surge flowrate of approx. 60,000kg/hr. So, now the anti-surge valve is open for bypassing some flow into the inlet. As you know, the operation of the compressor below surge point does not change the speed of compressor. However, whenever we have tried to ramp-up the flowrate upto 40,000~45,000 kg/hr, the rotor vibration has been increased upto s/d point triggering the whole unit s/d.

We are now investigating the source of the vibration with the vendor and the monitoring equipment. We can't find the reason yet. In addition, the lube oil leak happened at the coupling between the g/t and compressor.

Can anyone give a comment on the above??

Sorry for the poor english.

Regards,

Bae, K.K.

 

[TPL]

You really need to understand the nature of the vibration that causes the trip, in particular, what is the frequency (or frequencies)that are causing the problem.
Does the vibration increase with speed or does it build up whilst the unit is at a steady speed/load?
I would suggest you need a good vibration survey to include the following:

a) Start up from cold, load until trip occurs. Collect vibration spectra, orbits,1X amplitude and phase, and gap volts during startup, trip and consequent rundown

b) Upon coming to rest, immediately repeat a), to check for repeatability (or lack of) at all points in the test.

Here are some guesses (and they are only guesses) as to what you could be seeing:

i) As you increase load and the compressor speed increases, its approaches a critical speed (could be 1st or 2nd), increasing the rotot sensitivity to unbalance and generating high vibration.

You would need to look at Bode and Polar plots to test this hypothesis.


ii) Its a new compressor so there is an increased likelihood of a seal rub while the unit 'beds in'

You need to check repeatability of orbits polar plots and bode plots to check for this.

iii) Some sort of thermal influence develops on the compressor rotor as it warms through

 

[bk8548]

Dear TPL,

Let me briefly explain. Under the surge flowrate of 60,000 kg/hr, there is no increase of the compressor speed actually and the speed is steady, because the flowrate is controlled using the antisurge valve below the surge point.

And the antisurge valve operates automatically preventing the surge.

Of course, our vibration engineer is monitoring all the data and analyzing the Bode and Polar plot.

Thank you for your good comments, anyway.

 

[TPL]

Sorry - I should have read your post more carefully.

So....essentially all you are doing is closing the recycle valve and sending a greater proportion of the gas down the discharge line with the compressor speed remaining constant.

And, at constant speed, vibration increases to trip level.

My best guess is a rub. I would suggest checking shaft centreline position changes (derived from gap volts), looking at orbits and checking vibration spectra.

I would also suggest monitoring 1X amplitude and phase from low delivery conditions whilst the recycle valve is slowly closed until almost at the point of vibration trip and then opening up the recycle flow to see if the vibration levels reduce.
Are there any thrust position changes associated with closing the recycle valve?

What is the compressor critical speed compared to the operating speed at the deliveries you have quoted?

 

[arbainmahmood]

Dear Bae,

From which vib that tripped your machine - radial vib probe at which brg (x & y-), or axial displacement (z-) or casing vib when ramping-up flowrate?

What is the min flow? and this min flow should be called min flow to surge. Compr shud always run above min flow to avoid surging (to the left of compr curve) but you keep calling 'flow under surge 60,000kg/h'.

All you do is keep the recylce valve in auto mode & ramp open the discharge valve to get the flow, right?