Part 2 Sample generation

In order to conduct Reliability Analysis we require a sample. What is a sample?

Assume that this vertical arrow represents the time line on which maintenance work orders are issued and executed.

We’ll consider five work orders conducted over a period of calendar time.

Work order 1 was a functional failure of failure mode 15.

Work order 2 was a functional failure of failure mode 16, and so on.

On Work order 4 a maintenance technician suspended the life of failure mode 15.

And on work order 5 we caught the failure mode 15 while it was still in its potential failure state.

But this format in which the data is stored in the EAM, won’t allow us to perform Reliability Analysis directly.

We must extract a sample from the EAM’s work order database. Then, we may perform Reliability Analysis on the sample.

We’ll transform the work order data into an Events table.

Note that each work order generates two events. The failure mode life’s ending, designated by E and the failure mode’s life renewal or beginning event designated by the symbol B.

Failure Mode 15’s ending event by failure is represented by the symbol EF15. It’s renewal or beginning event is B15. Failure Mode 16’s ending by suspension is ES15 and so on.

Notice that in work order 5, for purposes of Reliability Analysis, a potential failure transforms simply to a failure. That’s because, although we may have avoided the direst consequences of failure, we haven’t prevented the failure itself.

The solid arcs join the beginning event on one work order to the ending event on a subsequent work order. For example an arc joins B15 to ES15. Each arc represents a life time of its respective failure mode.

We define a sample window by specifying two points in calendar time.

The arcs that appear within the window constitute our sample. Reliability analysis is counting the arcs in the sample. For a particular failure mode the number of arcs divided by the sum of all its lifetimes is approximately equal to the failure mode’s MTBF, or mean time between failure.

What about those partial dashed arcs, whose beginnings or endings lie outside the sample window? They are known as suspended or censored data.

The arcs that begin inside the window but end outside the window are called “right suspensions”. The arcs that begin outside the window but end inside are called “left suspensions”. The software enabled procedure that I will demonstrate will account for this suspended data appropriately.

## Post Tree Navigation

LRCM

RCM

Reliability Analysis

CBM