High Vibration In Gear box, multiple of TS but not GMF

Looking for some assistance on a high speed gearbox vibration problem. I have a attached several files of and the one labeled vibration history will give a overview along with some data. The gearbox amplitude reached .72 ins and 159G's peak to peak. The frequency appears to be a multiple of TS but does not line up to anything, and the dominate frequency changes from month to month. From 3414hz, 3596hz, 3775hz, 4133hz, 4495hz,. Fault Frequencies are below.

Motor trend is stable and low, compressor trend follows the gearbox trend, but is much lower, and both are in the attached document.

The gearbox is old and only has one proximity, which is on the high speed output shaft and the reading is always low, and has never been above 2mil

Important Frequencies

Input 29.8 hz , 4jaw coupling 119.2 hz, lube gear 417 hz  

Output 179.8 hz, High Speed coupling teeth 7903 hz, thrust pad 1437 hz

Gearmesh 6466 hz

  • Important numbers Gear has 217 teeth, pinion has 36 teeth, high speed coupling has 44 teeth, high speed thrust bearing has 8 pads per active and 8 pad per inactive, lube oil pump has 14 drive and 14 driven teeth,4 jaw love joy coupling all radial bearings are plain babbit and not tilt pad. Compressor vanes are unknown, but looking at the pictures, it appears there are around 17. Compressor has tilt pad bearings with 4 pads each and 8 pads per side of the thrust bearing.

Jimmy V.

Attachments

Original Post

My initial inclination, and based on the 1st attached document, is to get a properly mounted accelerometer on the casing so that you can insure you have valid data.  In the 1st attached document, you state "Route data was taken with a 100mv/g accelerometer, High frequency data up to 25,000 hz was taken with a high frequency 10mv/g accelerometer and wax mounted."  

At higher frequencies it is absolutely essential that the accelerometer be stud mounted, preferably in accordance with API 670.  Wax mounted just won't work due to the drastic downward shift in the mounted resonance of the accelerometer.

 

Thanks John,

I have stud mounted four of the 100mv/g accelerometers to the gearbox and collected data, and very close to the same as the rare earth flat magnet data. I need to purchase some high frequency 10mv/g accelerometers  and mount them to the gearbox as well. I have changed analyzers, and dumped it into a different software.

I have put a wireless vibration monitoring system on the gearbox as well (Wi-care), however it amped the sensors out.

I will put the stud mounted 10mv/g sensors on as soon as I receive them. 

Since the 100mv/g readings were stud mounted, do you believe that this is good data?

I will be at the plant Monday to take some pictures of where they are mounted.

Thanks again

Jimmy V. posted:

Thanks John,

I have stud mounted four of the 100mv/g accelerometers to the gearbox and collected data, and very close to the same as the rare earth flat magnet data. I need to purchase some high frequency 10mv/g accelerometers  and mount them to the gearbox as well. I have changed analyzers, and dumped it into a different software.

I have put a wireless vibration monitoring system on the gearbox as well (Wi-care), however it amped the sensors out.

I will put the stud mounted 10mv/g sensors on as soon as I receive them. 

Since the 100mv/g readings were stud mounted, do you believe that this is good data?

I will be at the plant Monday to take some pictures of where they are mounted.

Thanks again

I don't know what your system is but 159 g's peak to peak (are you sure pk-pk) is extremely high.  Even if 0-pk, the normal unit of measurement, 80 g's would likely be a transducer in saturation and thus your move to a 10 mv/g or even a 25 mv'g is justified.

Where are you mounting the accelerometers?  They should be mounted on the side plates (if fabricated housing) or the bearing blocks just above the bearings.  The spec limit of API 613 is 4 g's pk and the transducers are mounted horizontally.  

As a point of interest, 20 years of my career were directly involved with gearing while with a "premier" manufacturer of high speed gearing in the USA.  The highest sustained level I ever saw attributable to an actual defect was about 40 g's peak.  

 

 

3596 Hz / 179.8 Hz = 20.00 so problem frequency is 20 x output SS. Perhaps there are really 20 compressor rotor/diffuser vanes and not 17? The problem frequency is apparently shifting, but perhaps due to FFT resolution and load variation that causes motor speed to vary slightly. A typical 100 mv/g accelerometer is rated for 80-g and some only 50-g, but can be further limited by ICP voltage. I did "try" to measure 80-g on a gearbox with AF bearings, but the accelerometer with flat rare earth magnet was sliding down on a vertical gear case surface and moving horizontal on the gearbox steel support beams! The gearbox was found to have EDM on bearings, like your photos. 

What was done to prevent future shaft current, since motor bearings were trash, and gearbox bearings had indication of shaft current damage?

I would scan all surfaces (including both shaft couplings) with a microphone (audible and ultrasound), since this machine must be screaming for attention! 

Walt

Walt Strong posted:

3596 Hz / 179.8 Hz = 20.00 so problem frequency is 20 x output SS. Perhaps there are really 20 compressor rotor/diffuser vanes and not 17? The problem frequency is apparently shifting, but perhaps due to FFT resolution and load variation that causes motor speed to vary slightly. A typical 100 mv/g accelerometer is rated for 80-g and some only 50-g, but can be further limited by ICP voltage. I did "try" to measure 80-g on a gearbox with AF bearings, but the accelerometer with flat rare earth magnet was sliding down on a vertical gear case surface and moving horizontal on the gearbox steel support beams! The gearbox was found to have EDM on bearings, like your photos. 

What was done to prevent future shaft current, since motor bearings were trash, and gearbox bearings had indication of shaft current damage?

I would scan all surfaces (including both shaft couplings) with a microphone (audible and ultrasound), since this machine must be screaming for attention! 

Walt

To meet API 670, I think you will find that the accelerometers will be 50 mv/g.  I know that to be true for BN/GE, Metrix, CTC and most of the PCB line.  At 75 g's this will most likely place the accelerometer in saturation making most high frequency amplitudes questionalble.

John,

Maybe I am saying in wrong, If you look in the first document page 11, the is a waveform and the  impacting I am speaking of is the PK+ plus the PK-, and I am calling this peak to peak. Today we used  a 2140, and stud mounted 10mv/g high frequency CSI A022 2HF accelerometer .  The frequency's that are in question are around 3,000  and 4000 hz. When compared, the stud mount 100 mv, and the 10mv data looks the same. We also used a 100 mv with a rare earth magnet to take roaming measurements on the bottom of the bearing housing, and end plates. The data on the bottom was slightly higher than the top side readings. Peak to Peak impacting in the waveform was around 80Gs today. We also laid a flat washer on the top of the gearbox inspection and it was bouncing around.

Does this rule out it being "bad data" or is there another test I should do?

What did you think about the gear wear  pictures on the last page of the first document?

 

Walt,

I am not sure about the vanes being 20 in the compressor, but the amplitudes in the compressor are significantly lower, and I do see a 17X peak in the compressor reading.

A new motor was installed with properly shielded bearing in 2014. The gearbox has not been opened, other than the inspection cover since 2014, so I am not sure what those bearings look like.

I will scan the machine with our ultrasound machine this afternoon, thanks for the idea.

 

 

 

Attachments

Photos (4)

The vibration data looks okay; just very high. I once had a gearbox produce high modulation at 1x output shaft. A careful visual inspection of the full bull gear rotation revealed a few pieces of metal that were jammed/embedded into a few of the output gear teeth. Each time the bull gear passed the pinion the metal debris would cause a sharp impact. I postulated that the sharp impacts were exciting one of the bull gear vibration modes. Removal of the metal and "cleaning" teeth surfaces significantly reduced the vibrations. I not implying your situation is the same; just telling a story!

Walt

Add Reply

×
×
×
×