On August 10, 2005 we had a screw conveyor go down due to lack of lubrication in a SKF6012 2Z bearing. This is a motor mounting bracket bearing that is separated from the oil sump of the gearbox by a lip seal. I use vibration analysis to prevent these unscheduled breakdowns but I missed this one.
On July 12 my normal route vibration data did not trigger any alarms. On this equipment I monitor 9 parameters from Overall to PK-PK waveform.
On July 9 I was asked to collect vibration reading on this screw because it had tripped out on the evening shift the day before. The results of these readings were in the normal range except for the vHFD was over 7 G’s. The waveform amplitude was 6.18 g’s PK-PK. This was about a 10% increase but nothing to start running around about. The spectrum did not show any high vibration peaks. The highest peak in velocity was .04 at one order. The highest peak in the higher frequencies was .2 g’s at 2264 hertz. I decided I needed more data so I doubled my F-max to 5000hertz and collected another spectrum.
Now this spectrum showed me why I had a vHFD alarm. Now my PK-PK was 62.67 g’s PK-PK. My spectrum in acceleration was .86g’s at 3198.3 hertz. But nothing matches any bearing frequencies. And the bearing locked up the next day before I could collect any more readings and make the call to change out this bearing.
My questions are:
Is this the bearing’s natural frequency?
Should any bearing frequencies be present?
There is a lot of non-synchronous data but non matches this bearing.
Original Post