In a centrifugal compressor, a rise in vibration at Vane Pass Frequency (VPF) is noted. This is a multi-stage centrifugal compressor and is a part of compressor train driven by two turbines.

1- The vibration at 17 X (vanes of first impeller)

2- Vibration at VPF increased from 2 to 7 mm/s.
There has always been vibration at this frequency; however it never increased beyond 2 mm/s in the past.

3- Vibration measured one day before this data indicated no change from the historical trend.

4- No notable process condition change observed on this compressor.
5- This sudden change at VPF is a matter of concern as this machine was maintained few days earlier for thrust bearing damage. That incident occurred instantly and the machine tripped on high axial thrust security without any warning or process condition change.
Any Ideas?

The thrust bearing repairs may have incorrectly positioned the rotor, so blade-vane gap was reduced and raised pulsation amplitude. Another possibility is that the repairs affected a natural frequeny close vane-pass frequency, and now there is more resonant amplification.

Walt

Walt,
Data collection after repair and at normal operating conditions didn’t indicate any change in vibration values. This vibration increased a couple of days after repair job. So incorrect positioning of the rotor has been ruled out.

What about the inlet guide vanes? Where all the linkages put back in proper orientation? I would assume you a taking case readings, any prox probe data showing this rise also? Do you have some more info on the type of machine?
Dave

maybe now after some bedding in time something related to the repair has moved creating the situation walt has put forward? Are these high readings repeatable?

what was the rootcause for the thrust bearing damage in the first place? the rootcause may have not been addressed. think you have prox axial monitoring since syn gas train is critical 1 eqpt. what is the change like?

ingression of moisture is one of the possibility that may elevate 1st impeller VPF and moisture ingression is an enemy of balancing line since it is normally connected to 1st suction. you will have sudden axial jump if you have this phenomenon. from the start-up after the repair, the moisture maybe building up from the upstream that only after a few days you noticed it through axial change. i am trying to relate the initial TB failure to your current issue.

Dave, Jonesy

There are no inlet guide vanes and associated linkages. Proximity data is not indicating any change in the condition of the machine after repair. One reason for this could be that proximity probes are less sensitive to high frequency activities. Moreover, the data is repeatable.
Valve,
Root Cause of TB failure is being investigated. They have noticed gradual change in the thrust (proximity monitor), 32 seconds prior to its drift into danger limit, but no relation to process condition could be established. Interestingly, they have noted slight change in turbine (coupled with this compressor) rotor thrust exactly 32 seconds prior to this failure. They are trying to establish the relation.
One change noted after TB failure is abnormal steam leakage through end seals of the turbine. I think this has caused increase in temperatures at compressor pedestals and coupling covers giving rise to stiffness in the axial orientation. This in turn could affect the axial resonance and increase the casing vibrations at VPF. I have recommended, providing insulated protection against steam leaking from the turbine side to see the effects on the axial vibration response of the compressor.

Ali, if there is no nitrogen purge to prevent ingresion of leak steam into the turbine bearing box, your inhouse lab would have moisture contaminated oil sample from the reservoir.

I do have experience wrt excessive leak steam from tubine ends and it is represented in excessive water content in oil. We had to centrifuge the 22K litre reservoir on daily basis for 8 hours with significant water being collected i.e ~20 litre per day. In my case, we do not see changes in axial at the compressor.

I would infrared the comrpessor DE padestal skin temperature and compare with the rest which not exposed. Most of the leak steam from the turbine end seals would have dispersed outwards as it hit the turbine bearing housing.

Btw, when you guys stripped off the compressor for the TB replacement, did you see any sign of pitting due to moisture on the 1st impeller or any other abnormal findings? i still believe 2 events are related.

This message has been edited. Last edited by: Valve,

Valve,

They are preparing for theromgraphy to identify exact leakage point. I agree that such high leakage would give rise to moisture in the Lube oil system, which apparently is not the case here.

Only the TB was repaired by replacing the active side pads, the compressor was not completely disassembled.

Ali, for the leak steam check, a loose fabric tied to a stick and head it at the end of the turbine end seals is good enough imo.

No rise in moisture? If you do not have nitrogen purge or any means to prevent leak steam from hitting directly to bearing box, i believe you do not have significant leak of steam from the end seals. Considering other possibilities eg. surging, oil failure etc has been ruled out, failure of thrust balancing pressure via balancing line is what i can think of for your TB failure if the active you said is the side where the balance drum recieve the suction pressure. You might want to check your comp suction pressure and temperature (few data at different time) the day of the trip and compare it against FEW MONTHS before data. Same load condition.

Finally, the steam leakage point was identified. This was insulated steam inlet flange, the leakage was fixed by simple bolt tightening.

BUT no change in 17 X axial component observed.

Ali, on the leak, you are lucky that the leak is gone after you tightened it. For ring joint gasket (i think you inlet flange gasket is of RTJ), if the leak is significant and been for considerable period, steam cut would have occurred of which tightening will not stop the leak. Gasket replacement is needed. Good luck in indentifying the 17X exciter and hope you will not experience TB failure recurrence. Btw, it is wise for you to inspect the compressor impeller once the oppurtunity shutdown arises

This message has been edited. Last edited by: Valve,

Valve, The gasket, luckily in this case is "lens type" and most probable reason could be less than required tightening or bolts aging coupled with an insulation covering every thing including bolts.