Reliability Centred Maintenance · Chapter 5

Reading the Results

How to read the outputs and what they let you decide.

The output is a maintenance programme with a reason attached to every line, and the reasons are worth more than the programme. A programme can be copied; the reasons are what let the next team change it safely.

1. The task requirements

The deliverable, on the worked example: five scheduled tasks and five recorded decisions to do nothing. Two of the tasks are new work, which is what goes to task analysis next.
The deliverable, on the worked example: five scheduled tasks and five recorded decisions to do nothing. Two of the tasks are new work, which is what goes to task analysis next.
FieldWhy it travels with the task
The task, and what it is meant to detect or preventA task without its failure mode cannot be reassessed
The intervalThe number the maintenance plan runs on
The computed maximum, where one existsSo packaging decisions stay visible and defensible
The consequence categoryWhich effectiveness test it passed, and how hard the bar was
The inputs the interval depended onThe P-F interval, the rates, the tolerable multiple-failure rate

That third row is the one most often lost. An interval adopted at 500 hours against a computed maximum of 952 has a factor of 1.9 in hand; the same 500 hours recorded without the maximum looks arbitrary and gets extended by somebody who does not know what it was protecting.

2. The decisions to do nothing

Run-to-failure entries are results, not gaps. Each one records the mode, the consequence category, and the fact that no task was both applicable and effective. On the worked example five of nine modes end here, which is normal and is the reason the programme is affordable.

3. Design requirements

Every mode where the consequence is safety or environmental and no task can be found is a design output, not a maintenance output. So is every case where the arithmetic demands an interval nobody can deliver. These go to the design authority as requirements with the analysis behind them, and they are the highest-value thing an early RCM analysis produces.

4. What the programme costs

Frequency times duration, per task, rolled up. On the worked example the analysis added 44.7 maintenance man-hours a year to a system that was carrying 164.9, a 27 per cent increase, and it can defend every hour of it. The task analysis turns those requirements into hours, people and parts; this module supplies the requirement and the interval.

5. The revision plan

TriggerWhat it invalidates
Observed failure rate differs from the assumed oneEvery computed interval that used it
A P-F interval turns out to be shorter than assumedThe condition-based task built on it, immediately
Age exploration produces life dataThe intervals that were set conservatively for lack of it
The operating context changesPotentially the consequence categories, and therefore the answers
A design changeThe mode list, and everything downstream of it

What the results do not support

  • They are not a schedule. Tasks and intervals are inputs to maintenance planning, which packages them into checks against opportunity, access and downtime.
  • They are not an availability model. RCM says what work exists; whether the system meets an availability target needs the corrective work, the repair times and the logistics as well.
  • They do not make a case for more maintenance. A rigorous analysis usually removes tasks. A result that adds work everywhere is a signal that the effectiveness tests were skipped.
  • They are only as good as the mode list. The analysis cannot decide about a failure mode nobody wrote down, which is why it is built on the FMECA rather than on a workshop's memory.
  • They expire. Intervals derived from assumed rates are provisional until service data arrives, and an analysis never revisited is a record of what was believed on the day it was signed.

Want to see this on a live system model? Request a walkthrough.