Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
electricpete

Sr. Member
Registered:
Posts: 651
Reply with quote  #1 
I've been asked to chase some hypothetical  scenarios about setting bearing housing velocity shutdown limits on rolling bearing machines when there is a critical need to run a machine that already has an identified high vibration.

There are obviously a wide variety of factors I would consider in this situation. Characteristics of the vibration, Trend of vibration, Monitoring strategy, Plant Consequence of failure, personnel hazards of failure, consequence of shutting down the machine, vibration and failure history of similar machines,  risk tolerance, etc.  I have a reasonable enough handle on those aspects for the situation of interest and I'm not looking for general advice.   I want to focus exclusively on a question related to "industry" documents.  (this will form only a portion of my decision making process).

QUESTION: What is the highest velocity limit you have seen mentioned in an industry document (*) as either a vibration limit, severity guideline (example imminent failure), or a shutdown criterion?  Let's say the equipment in question is 1800rpm.  

* industry document for my purposes could also include books, articles, even websites.  Credibility of the reference can be judged separately.  

I've found a reference mentioning 0.7ips (overall "pk/0", which would correspond to 12.6 mm/s rms)... has anyone seen any higher number mentioned ?







Beatnik

Sr. Member / Supporter
Registered:
Posts: 183
Reply with quote  #2 
https://www.irc.wisc.edu/export.php?ID=158

In this doc they talk about a 0.15 IPS RMS limits at bearing frequencies. That make sense to me.


Good luck, it's really not an easy task, specially if you are limited to overall velocity!

John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #3 
Very specialized application; gas turbine driving through an epicyclic, planetary arrangement intended for a generator drive.  The individual setting the unloaded test acceptance level of 1.0 in/sec peak happened to be an employee of the GT OEM but also a gear expert.  That's the highest I've ever seen as far as a spec, and again very specialized machine and solely on the test stand recognizing components are moving around in a planetary without load.   
electricpete

Sr. Member
Registered:
Posts: 651
Reply with quote  #4 
Thanks. 

Interesting special case of 1.0ips, sounds like it was well thought out and very specific to the situation.

Interesting link about rolling bearngs. I'm not looking at rolling bearing defect. I do think it's a pretty bold limit of 0.15ips rolling bearing defect frequency as limit for "acceptable" on page 41....  I would certainly call it unacceptable long before that although shutdown recommendation would depend on circumstances.  But I am focusing on overall.  In the same document you linked on page 36they have a chart where they identify 0.63 ips overall as "very rough". 

I did find this link below which identifies 1.0 ips overall as level "extreme"
http://www.azimadli.com/images/severitychart.jpg


Beatnik

Sr. Member / Supporter
Registered:
Posts: 183
Reply with quote  #5 
There is of course ISO 10816-3 that is used a lot.


Here is limits determined by the canadian government:
http://www.oocities.org/indouae/Technical/page3v.html
Curran919

Avatar / Picture

Sr. Member / Supporter
Registered:
Posts: 599
Reply with quote  #6 
There are plenty of super specific standards that are applicable technically to roller bearing machines. Single vane submerged waste water pumps running outside the POR and on 'clipped in' foundations can get up to 16mm/s allowables.

I don't see how this is going to help with your machine though...
DCD

Member
Registered:
Posts: 4
Reply with quote  #7 
When we see 1.0 IPS the next warning is usually smoke (or breakage) so I tell them to monitor temperature hourly looking for an increase, when they see a increase they may be minutes from catastrophic.
electricpete

Sr. Member
Registered:
Posts: 651
Reply with quote  #8 

The task that motivated this question is complete.
Sorry for seeming a little narrow about the question, it's just not practical to explain the whole context.

Thanks for all the input.  It was helpful.

Vibe-Rater

Avatar / Picture

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,398
Reply with quote  #9 
i don't like using the charts which inevitably are overall measurements.  0 - 1000Hz  0r 0 - 10kHz. And you all know, the contributions from imbalance so heavily effect the numbers velocity.  So you spend your life balancing and improving structural stiffness getting that 1x down. When a bearing goes it hardly makes its mark on overall. I work mostly with paper machines so the important things happen between 0.3x and up to about 150x regardless of running speed.  I have an example where I found a bearing fault and out of curiosity looked out to 300x and the region between 150x - 300x really was not of any use. From a past employment I met some people who were running around spruiking 0 - 10000 Hz.  Again I adopted that at a particular site and found it of very little use.

So, about 6 or 7 years ago. i get a phone call from a lawyer. Can you please go and check out a bearing?  Smelling a rat straight away I prefer to avoid all contact with lawyers. So they insisted and I visited site. Immediately see that another company was there before. I pry it out of them and they gave me the report. I looked at the data displayed and immediately saw a bearing fault. But they made their decision on the 10 kHz spectrum and called it all OK.  And the bearing failed catastrophically the day after measurements were taken.  And is why a lawyer was involved.

I was nice about it. You know go down the road of - well no prior / historical measurements with a bit of black magic thrown in. I basically did not want to get embroiled in anything like that.

The fastest shaft I deal with is normally 3000 rpm (Australia 50 Hz grid) but there is one I can think off 7600 rpm or there abouts.  Tap into the Bently sensors for that one. So normally I would look at an Fmax of 7500 Hz when encountering 2 pole shafts. And the rest is a lot lower.  And then there's the question of useable Fmax from a flat earth magnet mount. I know the response graphs probably mostly agree on about 5Khz but I reckon with the right approach you can reach 7.5K. So all good.  Happy hunting - for machine problems. rgds
Tomasz Malcher

Member
Registered:
Posts: 3
Reply with quote  #10 
1 inch/sec RMS (10-1000 Hz) . Obligatory shutdown limit for reciprocating compressor cylinder-end vibration. Major gas compression operator guideline.
John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #11 
The highest spec limit is one thing, but just how high have you measured?  

I was involved with a newly installed gearbox that was reported to have high vibration.  I went to the site, magnetically mounted velocity coils all over the gearbox and was sitting off to the side when they started the machine.  I recorded everything to tape but almost immediately commented to the facility personnel that the gearbox was very smooth.  They informed me that almost all the transducers had fallen off due to the vibration!  After stud mounting and some stronger magnets, we established that the level was approximately 5 in/sec, predominately at 1800 cpm which was 2X of the diesel prime mover.  The issue was a torsional mode around 1770 cpm coupled with a mounting base resonance of the gearbox support structure at about 1750 cpm.  
Danny Harvey

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,403
Reply with quote  #12 
I cover an eddy current separator that I regularly see 4-5 ips at about 3000 cpm.  I don't know how much longer that will be "normal" though. [wink]

At the same site there is a vibrating fluid bed separator that operates at about 10 ips at 818 cpm.  We are about to mount accels on the drive bearings so that I can do predictive maintenance on them.  I routinely track orbits on the resilient mounts at the same 10 ips.

I had one case where I couldn't get an accel to hold still on the gearbox housing of a T-G drive.  The graph I got was a ski slope with about 60 ips at 0. It was an obvious overload.  

Using just PeakVue I was able to determine that the new low speed gear that had been cut actually had three high points instead of being round.
spciesla

Sr. Member
Registered:
Posts: 107
Reply with quote  #13 
I had a diesel engine driven firewater pump vibrating around 2 ips.  My engineering manager at the time called BS, saying that the magnet mounted accelerometer would have fallen off.  I told him, "You're right, I had to hold the base of the magnet with my hand to get the reading!"
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.