Hi guys, I would like to tell you that I know what is wrong with this machine but… I don’t know… The true is that I’m a beginner analyst and we are implementing a Pdm program in our plant and I think I need a hand analyzing this problem. The whole machine is a lamination train drive by a 600 HP motor, 1180 RPM, connected to an Inertia wheel (see the picture attached) connected to a reduction gearbox (1180 to 180 RPM). The thing is that overall velocity vibration in all directions in both bearings of the Inertia Wheel are really high with 1X, 2X and 3X RPM peaks, I think indicating some misalignment combined with unbalance( See the spectrums attached). However, overall vibration of the motor and gearbox are not even close to the Inertia wheel vibration amplitudes. On the other hand a strange thing is the peakvue measurement on the inboard bearing of the Inertia wheel where I got 37.76 g’s on the waveform’s peak to peak parameter and high vibration amplitudes on the 1X and 2X RPM frequencies on the Peak vue spectrum (see attached figure). The thing is I don’t know what to tell about 1x and 2x RPM peaks in a Peakvue spectrum, I mean there’s no even a little peak on any of the bearing fault frequencies (SKF 22332CK/W33)… As you can see I really need some HELP here…..

Other Thing I want to ask you is that when I measure the low speed side of the reduction gearbox (180 RPM), I get all the significance peaks from the high speed side (1180 RPM) meaning that I never going to be able to find out problems related to 1X, 2X, 3X….RPM frequencies on the low speed side. What can I do here? I’m using an all purpose 100mv/g accelerometer with a two pole magnetic mounting…

THANK YOU VERY MUCH FOR YOUR TIME….
AND PLEASE EXCUSE MY ENGLISH…

Jose Luis

Spectra,_Inertia_wheel_Picture.doc (989 Kb, 48 downloads) Spectra, Inertia wheel picture

Jose,

The main concern here to me is high amplitude impacts in PeakVue. First, I suggest using Autocorrelation function to get a better picture of impact periodicity in the TWF. It may show presence of another frequency modulated by 1x obscured in the TWF and spectrum. Does PeakVue look this way only on this measurement point (I guess it is IB of the wheel) only? Could be looseness somewhere in the wheel or bearing fit.

For low frequency measurements, such as 180 RPM, set the low cut off value lower and use 500 mv/g accel.

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

David,

Thanks for taking the time to review this problem... I don´t have the autocorellation option on my software for TWF... The spectrum on the file is the inboard one. The outboard peakvue spectrum has only "noise" raising the overall up to 3.29 G´s but it doesn´t have any high peak raising over the noise. However the peak to peak measurment is 15.78 G´s....

Jose Luis

Jose,

Please post OB bearing data: peakVue and normal.
Make sure PeakVue is taken with consistency as its peak values are critical to accel attachment.

David here you have the other spectra as you request...

Jose

spectra,_outboard_bearing.doc (296 Kb, 25 downloads) Outboard bearing spectra

Jose,
I would not have called IB and OB readings significantly different considering probable inconsistency in PeakVue (PV) data acquisition. In my view there is a possibility of looseness and consiquent impacting in the wheel fit on the shaft sensed on both bearing housings. The TWF in PV shows a lot of randommness as oppose to sharp repetitative impacting at 1x which is consistent with this diagnosis.

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

José, two things :
1.- For low frequency diagnostic is useful spectrum in mm/sec wich let you to see clean and clear time wave form like shown in your spectrum, so you need to zoom the spectrum in lowest kcpm intervals ( you showed each 20000 cpm).
2.- About the diagnostic, I am agree with Dave, I believe in loseness, common in flywheels. About the solution of this kind of problems you can think in axial preload using rubber attachment in the coupling. Some friend in this forum are experienced, any thought are welcome.

Regards and have a nice day.

Looks like input gear damage, maybe cracked tooth or severe wear? Where is this reading taken? Use flat magnet on flat surface if possible. Do you have a picture of the machine? Zoom in on the time waveform and see if the spacing=GMF. Can you visually inspect?