One of the biggest challenges in condition based maintenance is answering the big question of production managers when you inform them of a detection of a failure mode.
When a PdM technician tells a manager, I can see that we have a problem in this bearing, or something similar, the question that is most often asked is : “When will it fail?” or “How long do I have left”.
Answering this question usually involves the intuition of the Maintenance PdM technician, how well does he know the machine, how well is he familiar to the operational environment, are there new operators on the machine, will there be new operators on the machine, will there be a change in the maintenance procedures affecting the failure mode that was detected by the condition monitoring… and the list could go on and on… and on… and on…
The period from the P-F points on the failure curve are dependent on too many immeasurable variables to be able to say, for example : “You have one month before this bearing will break down!”.
It is often extremely tempting to express an opinion of how long the period will be from P-F on the failure curve, however there must always be a “grain of salt” with it, so that the production “side” is aware that this is not a guaranty at all.
Another point to consider is that once a failure mode is detected almost always there is an element of waste added in the process, more power is needed to do the same. There are leaks, internal or external, that are wasting energy, etc.
The point of all this here above is that as soon as a failure mode is detected steps should be taken to plan a repair of the failure. This is the foundation of precision maintenance.
When analyzing what went wrong, to prevent further failures from the same source, focus should be given to the Installation point, because before the P-F curve we have the I-P period, that is the Installation to Point of failure detection point… but more on that later.
I hope that you have enjoyed to read this post and will leave a comment.
Related articles
- How To Get Results From Maintenance Systems Montreal2012 (slideshare.net)
- Fixed and improved… better than new! (bjarniis.wordpress.com)
- Condition monitoring, is it a modern day Chrystal ball? (bjarniis.wordpress.com)
- Measuring what did not happen. (bjarniis.wordpress.com)
Bjarni Ellert Ísleifsson, CMRP 
Great introduction.
Thank you.
Hi Bjarni: We recognized this problem with bearing failures years ago and developed a system called Multiple Discriminant Analysis tm. to address it. If you send me your email address I’ll send a few papers that describe the system. I think you’ll find it interesting.
Thank you for your comments John. My E-mail address is bjarniis@gmail.com