Unable to detect vibration issues in slow RPM conveyor system

[Orga78]

Hi,

What is the best practice for vibration analysis technique or method to be able to detect bearing failure symptoms in slow speed conveyor belt system?

We recently had a bearing failure in one of our vertical chain conveyor system. Our route-based vibration measurement did not spot any anomalies months leading to the event.

Thanks,
J

 

[hacksaw]

What sort of bearing lube? Greased or oil reservoir.

 

[Orga78]

It's a greased bearing

 

[hacksaw]

If low rpm, perhaps IR temp, or accoustic survey (typically ultra sound) to filter out the equipment noise.

At a minimum keep a record of the failures to get some idea of the failure rate.

 

[Tmoose]

Is this the first ever bearing failure that escaped notice?

And what specific kind of bearing(s) ?

Please elaborate on the bearing "failure " . More pictures than words, please.

Are the route measurements made on the actual bearing housings, or some hunk o' frame etc several inches from the actual bearings ?

What kind of parameters is the "route-based vibration measurement " gathering and I assume trending.
Any high frequency measurements (over 1000 Hz to maybe 3000 Hz or higher, inclusive ) in gs?
What type of sensors, and how are they attached during route measurements?

 

[3DDave]

It might be there were no symptoms. I can picture a case where lubrication failure allows a tiny rate of wear of a bearing cage that progresses until the cage fails and the balls or rollers escape. It will be working fine right up to the first rolling element falling out of position and going to noticeable failure in one revolution of the bearing. There won't be any change in vibration until then and there won't be a detectable temperature rise.

Grease is exactly the lubrication system that can allow this to happen. The oil in the grease escapes and the thickener does nothing. If there is a forced-lube system, the thickener can block the entrance of new grease or divert new grease without purging the thickener. Oil lube is superior in avoiding that.

 

[electricpete]

good comments already.

envelope / demodulating methods are supposed to be good for detecting slow speed bearing defects.
They work to highlight the low-level periodic impact behavior.

Were you using demod/envelope techniques? Did you go back and look at the demod spectra for signs of fault frequencies (and twf for "magnitude") ?
If you were using demod and it didn't show anything, you might look at some of these possible factors:
[ul]
[li]The conditions of attachment of the accelerometer to the machine can also attentuate the high frequency components that demod techniques may rely on.[/li]
[li]the location of monitoring (how close to the bearing) that Tmoose mentioned.[/li]
[li]the particular user-selected parameters of the setup of the demod can be very important.[/li]
[/ul]

see abstract here

https://www.vi-institute.org/publications/slow-speed-bearing-fault/

Also case study here

https://rms-reliability.com/wp-content/uploads/2019/05/RMS-Vibration-Case-3-Slow-Speed-Faults.pdf

If you google low speed bearing fault detection you'll see a lot of results relating to ultrasound monitoring. [walt] strong on this forum has used ultrasound monitoring.

We use demod (csi peakvue) to assist with our early stage bearing defect detection on normal speed equipment (the slowest machine I look at is 324rpm) and we confirm that looking at the "normal" spectra and twf. To be honest we don't have great performance with the demod, there are a lot of false alarms and a few real problems that didn't show up on peakvue. Our vibration engineer is tweaking the parameters. Although it doesn't seem to be perfect, personally I would still look at vibation demod first and try to understand why it didn't work before burdening the program with another technology (ultrasound).

=====================================
(2B)+(2B)' ?

 

[Orga78]

Hi All,

Very valuable comments and inputs indeed! Thank you. We have recently performed visual inspection of the failed bearing and found the internals were in low lubricant condition.
Suspected the bearing seal had damaged for quite some time, thus the grease pumped in through the housing port did not gain entry into the bearing inner, but outer.
Apparently, the PdM guy who took the vibration measurement only had horizontal direction set-up in his analyzer route database & setting for the conveyor. They include ENV (SKF) as well - but did not pick up any high unusual impacts (due to only H is being measured!) The FMax is 1000Hz.

Other technique to enhance PdM is also now considered, including ultrasound & IR temperature.

Inputs from 3DDave is almost on-point. However, the type of grease, its application and relube interval do not raise any red flag.