analyse the cause of vibration of generator bearings 1

[azimuddin]

I am in-charge of condition monitoring of our power and desalination plant since one year. I have ten years experience of maintenance but one year of cause diagnosis of vibration, which is not enough, however I have accepted the challenge.

When our organization purchased a vibroport-41 from Carl Schenck , Germany. A trainer came with the instrument and given us training in one week. How to use the instrument and how to diagnose the cause of vibration with FFT analysis. We are using velocity pickup VS 080.

I have found vibroport very useful device, but need your help to analyse the readings to come to some conclusion. Please study the following readings and let us know what is the cause of vibration.

Vibration of DE and NDE bearings of a generator (capacity 20 MW) in vertical direction are measured and recorded. The generator shaft is coupled with turbine shaft through a reduction gear. Turbine shaft runs at 4900 rpm and generator shaft at 3000 rpm. Observed that the readings are continuously varying in a zigzag pattern between a values of 6 to 22 mils pp. no single peak value sustained for a longer time. The graph of the vibration on time domain obtained as vertical lines of different size (joint by head to tail) side by side. Same situation observed on DE and NDE both.
FFT analysis reveals DE= 6 mils at 1.4Hz , and 1 mils at 49.9 Hz >>>>> NDE= 10 mils at 1.1 Hz and 0.7 mils at 49.9 Hz.

If a horizontal line drawn over the graph joining the mean values of each spikes it will be at a level of approximately 10 mils pp on both DE and NDE bearings.

The vibration range reduced to 4 to 14 mils pp when the generator started generating 5 MW. In this situation the average seems to be 6 mils on both DE and NDE bearings. At the same time vibration measured in rms velocity and found as 3.75 mm/s rms on DE and 3.5 mm/s rms on NDE bearing. No fluctuation in the readings observed in case of rms velocity.

Very strange thing happened. I simple removed the connection of input1 and input 2 from one vibroport and connected them to input 1 & 2 of the other vibroport and I got different readings of the same measuring point. Overall vibration reading with one vibration is fluctuating between a range of 4-14 mils while with the other it is between 3 –4 mils pp of the same measuring point with the same pickup and same set up.

When started the machine we found that the NDE bearing is vibrating very high. Later on both became same. Turbine vibration remains at 1.8 mils on compressor bearing whatever was the situation.

Now my question is
(1) what is the cause of vibration on the generator bearing
(2) why the vibration level (during FFT) is about 10% of the over all level of vibration at rotor speed at the same point. Is there any relationship of the vibration level at 1 x rpm with over all vibration.
(3) In our case we see overall is 10 mils but FFT analysis reveals only two peak values one as 6 mils at 1.4 Hz and other as 1 mils at 49.9 Hz. It is very clear that at 1 x rpm the vibration level is only 10% of the overall level. What is your comments here.
(4) If only the unbalance in a machine then what is the value of vibration at 1 x rpm when we take action to balance the machine.
(5) Vibration of 4 mm/s rms is acceptable or not?
(6) Why both of the vibroports are not showing same vibration readings. Whether Schenck has noticed it. What is the reason of it. Which one we will accept as true.
(7) Bases on the above readings and observation our generator shaft is balanced or not balanced. Any balancing required?
(8) If a machine is stabilized at its operating speed, lets say, we see no vibration fluctuating since last one hour but we see the phase angle is changing from 10 to 190 degree , then what is the reason.
(9) If the horizontal vibration is higher than the vertical vibration on the NDE bearing of the generator then what is the reason.
(10) What shall be our next step in the above case

 

[Franko]

Are you sure the machines are really vibrating and it is not instrumentation errors - especially at frequencies below 2 Hz? You should consider obtaining an accelerometer calibrator such as one offred by Endevco:

http://www.endevco.com/test/test_main.htm?news/28959F-port-accel-calib.html~mainFrame

. They also have information on making proper measurements, including correct mounting.
It is doubtful that the vibration near 1.4 Hz is real unless the entire foundation is vibrating at a very low frequency. The low frequency greatly affects displacement fluctuating readings, but much less for velocity readings since:
velocity (0 - pk)= 3.14 x displacement(pk-pk)x frequency. If you are using a piezoelectric accelerometer, there is often a lot of noise near DC. It is best to filter out this noise - below the rated frequency response of accelerometers. FFT analyzers should be set in the "AC coupled" - not "DC coupled" mode. If the instrument uses a velocity probe, it could have errors at low frequencies as well. The generator electrical field could be causing noise. We used to place an aluminum cover around old IRD velocity probes for line frequency noise.
If you can show that the low frequency data is in error, then true bearing housing vibration at 4 mm/sec (0-peak), measured in line with the bearings is considered generally acceptable by various vibration standards.

 

[dgallagher]

Franko makes some very good points. Suggest you also take a look at the various application notes on turbine supervsory instrumentation and vibration at:

http://www.reliabilitydirect.com/appnotes/appnotes.html