Some things reliability can learn from RCM

Some of the stuff that came out in the original RCM report has often been either maligned or just lumped in with RCM. Over and over again some of these principles have proven to be some of the underpinning concepts of reliability.

This is particularly true when I work with people who have deep personal experience in reliability, but have had no dealings with the information in this report.

Here are a few of them. I don't think this is going to change the world, but hopefully it will help you to sharpen your own focus a little and be better prepared for these discussions when they arise.



1) The over dependence on data. 

A continual issue and pone you see time and time again. The fundamental problem here is what Resnikov eluded to in his book Mathematical Concepts of Reliability-centered Maintenance. (Download)

Most historical analyses fail because the data is rarely available? And why is it rarely available? Yes there are issues with software and data capture systems. And yes there are work process issues.

But the fundamental reason that there is a lack of data is because there has to be.

Before you can have failure data you need... failures. And failures cost a lot of money, lost revenue, they kill and harm people and they damage the environment.

You need to be very well aware that the vast majority of decisions you will make will be made in the absence of data. And you need processes and techniques to deal with that.


2) The over dependence on criticality studies.


The important assets are not the ones that are going to help you make a quantum leap in performance. it's the ones that are bleeding !

3) A misunderstanding of how complex assets fail.


This is like a mass bathtub delusion. One that has had many papers written to try to support it, and many software products created trying to sustain the thinking further.

It is often hard to relate to issues of random failure. With many people preferment to think that it is only when a device doe not have a wear memory of some kind.

The critical piece of knowledge here is that complex assets do not have a dominant failure mode. Rail tracks will fail due to the wear between the wheel and the track in the vast majority of cases.

A bearing will not fail due to wear in the vast majority of cases. They have multiple failure stresses, leading to a constant or near constant failure characteristic.

4) Focusing on what an asset is not what is does.

This can be a very detrimental element. Particularly in large scale organizations with many similar assets.

This thinking is what leads to wholesale wallpapering of similar strategies across an entire asset register. Leading to over maintenance in a good situation, under maintenance in a bad situation, and missed failure modes in the worst case scenario.

Assets do not have a static list of failure modes. They will fail differently depending on how they are used.
____________________________

There are of course many others, but I had a few meetings in the past few days with some deep and skilled reliability people and these issues were ones that came up repeatedly. They are issues that seem to be wired into peoples DNA. Hope this helps.