All Member Open Forum

Hi everyone,

At the company I recently joined, a wireless tri-axial sensor based vibration program has been already started.

Many of the sensors are not mounted in 'true' horizontal of vertical direction- in my opinion, because of limitations of wall thickness of bearing housing/motor cover. (see a picture below; the other issue is the sensor on the NDE motor bearing- I think it is too far away from the bearing to be able to monitor the condition or any other problem like rotor unbalance)

The company responsible for the vibration monitoring (who also did the installation), argues that it does not make any difference for analysis, even in cases where one might want to compare vertical to horizontal readings (e.g. misalignment vs. unbalance- which of course can be also worked out from the ratio of 1X to 2X running speed and so forth). According to them, as long as the sensor provides the axial reading and one radial reading, this is enough.

Does anyone have experience with this or can guide me to some standard/theory to confirm or disprove the notion that sensors mounted 'askew' can still deliver the accurate diagnostic results?

I have done my bit of vibration analysis in the past, but I do not hold any certification.

Regards,

Ignacio Diaz Ducca

------------------------------
Ignacio Diaz Ducca
Maintenance Manager
McCain GmbH
Germany
------------------------------

Years ago I had a conversation about this with the Cat IV vibration analyst who owned the company that monitored our equipment. Using handheld data collectors, they were collecting one radial measurement at each bearing (rather than a horizontal and a vertical) and one axial measurement for each shaft. His argument was that, for monitoring purposes, one radial + axial is enough to tell you if there is a problem. Whether it's a bearing fault, misalignment, or whatever, it will show up. Then, on the occasions where a problem is detected, it may be necessary to take additional data for analysis purposes. There is already enough information to diagnose bearing faults, but true horizontal and vertical readings, and maybe phase readings, may be necessary to diagnose things like misalignment. I have to say I agree with him.

With your wireless installation, I have no doubt the inboard sensor will detect any problems that develop, as long as they at least aligned either the X or the Y axis with the shaft. The outboard sensor may or may not. I agree it's in a crappy location - they should have installed a fin adapter and put it closer to the bearing. I have been told a sensor will pick up a bearing fault if it is within 36", but I've never read that in a reference anywhere, and of course best practice is to mount it as close as possible. 

When it comes to analysis, I'm not so sure you have all the information you need. The whole point of separate horizontal and vertical measurements is that equipment will behave differently in those two directions because the stiffness of the mechanical system is different in those two directions. If you can only compare a 45 degree measurement to a -45 degree measurement, I suspect they would both look the same and you won't have as much information as you could have.

If this was my installation and I had a way to take more detailed follow-up readings after an anomaly is detected, I would say I was good to go, at least with that inboard sensor. If I was relying solely on this company's data to make maintenance decisions, I would make them move both of them to better locations. (My opinion only. I'm curious what others have to say.)

------------------------------
Dale Nicholson, PE, CMRP, CRL
Reliability Engineering Mgr
Evonik Corp
------------------------------

When mounting fixed sensors you should endeavour to optimise the transmission pathway from the target bearing to the sensor. Frequently this involves making a judgement call regarding the internal structure of the unit being monitored as diagrams can be inaccurate or misleading.

My method for optimisation is to employ a mobile sensor to ascertain exactly where the signal appears strongest; that's probably the best position for mounting the fixed sensor. Where possible I always align the horizontal sensor axis with the shaft; that ensures at least one axis is correct; it's a useful principle with vertically mounted equipment.

Once the fixed sensors are applied you will always be collecting data from the same orthogonal axes even if they do not appear true externally.

Hope this helps…

------------------------------
DUNCAN FALLOW
CBM Engineer
AVT Reliability
Huntly
------------------------------