Mean Time Between Failure

Try [Average interval between failures] multiplied by [number of machines]? Is that simple enough?

We have data for about 5 plus years

Good.

How does that information help Anonymous Poster #1?

Looks great, will share results.

Thanks

"mean time between failure" I'm usually in the maintenance shop having coffee with the guys lessin we're out behind the spares shop puffin our Marlboro's

First you must decide how many modes of failure exist. Then you need to know the failure rate of each of the failure modes. After you have collected the data, there are a couple of ways to approach the problem. The simple way is to take the square root of the sum of the squares.

MTBF = sqrt((Failure Mode A)² + (Failure Mode B)² + ..... (Last Failure Mode)² )

Unfortunately this will only predict how much downtime each machine is likely to experience using current practices. If you want to do something about it, then you need to realize a couple of things. First, each machine is unique. This means that what ever happens to machine #1 has nothing to do with what is going on with machine #2 and so on, except for some catastrophic issue with shared power feeds.

Next, you need to understand that every failure mode has some unique characteristic that can be used to detect an approaching failure. But finding that characteristic may require tools, techniques, or statical methods that you do not have at this time. Acquiring this kind of information takes seasoned engineers who realize that nothing is simple. And therefore, there is no simple answer to your question.