The short answer
Your own removal history — your sectors, environment and utilisation — is a better predictor of what comes off next than a fleet-wide average.
The useful output is a planning forecast: which removals, parts and shop visits to expect, so they land in a gap in the schedule instead of the middle of one.
A forecast informs planning. It never changes airworthiness status, and it should read from your maintenance tracker, never write to it.
Why the average is the wrong number
Manufacturer and industry reliability figures are built across many operators. They are the right starting point for an inspection programme and the wrong number for planning next quarter. An operator flying short sectors in a hot, sandy environment removes some components far more often than one flying long legs in mild conditions, and the average hides the difference.
What you need to start
- Removal records: part number, serial number, tail, position, date, and whether the removal was scheduled or unscheduled.
- Utilisation: hours and cycles per tail, ideally per day.
- Shop findings where you have them, because a removal that finds no fault is a different signal from a confirmed failure.
- Context worth adding later: routes, climate, and operating patterns such as average sector length.
The data rarely arrives clean. Serial numbers are typed several ways, removals are recorded as free text, and the same event appears in the tracker and in a spreadsheet. Cleaning and joining it is usually most of the work, and it is worth doing even if no forecast follows.
What a good forecast looks like
A useful forecast is specific and humble. It says which components on which tails are likely to be removed over the next planning horizon, with a range rather than a single date, and it shows its working: the history and utilisation it is based on. It should be measured against what actually happened, continuously, so everyone can see how much to trust it.
From there the planning follows: parts demand against forecast removals, shop visits placed in a schedule gap, and a clear view of the MEL items you are carrying and what they cost you.
How this fits a reliability programme
Operators with a continuing analysis and surveillance system — required for Part 121 by § 121.373, and for Part 135 aircraft of ten or more passenger seats by § 135.431 — already analyse the performance of their maintenance programme. A removal forecast is a planning tool that can draw on the same data. It does not replace that system, change an approved programme or set an interval.
The line it must not cross
A forecast is a planning input. Airworthiness determinations belong to certificated people and approved processes. The way we build these tools, they read from the tracker and never write status back to it, and nothing they produce releases, grounds or signs off an aircraft.
A sensible first step
Take a year of unscheduled removals, however they are recorded, and measure whether a forecast built on your own history would have beaten the plan you actually used. If it would not have, you have learned that cheaply. If it would, you know exactly what the gain is before building anything.