Unscheduled defect planning: the manpower buffer your check schedule forgot
Every base check I have ever supervised had two plans: the one in the planning office, and the one rewritten after the first deferred defect, bird strike or leaking damper arrived. Unscheduled work is not a surprise category. It is a measurable workload with a distribution, and if your roster has no reserved capacity for it, the check package pays the bill.
01Stop pricing defects at the annual average
The common mistake is to take last year's unscheduled hours, divide by 52, and call the result a weekly allowance. That gives you the mean, but maintenance does not arrive at the mean. It arrives in clusters: three aircraft serviceable on Monday morning, one AOG and two MEL-critical defects by Thursday.
Use your maintenance information system to extract unscheduled man-hours by week for the last 12 months. Exclude scheduled inspection cards; include defect troubleshooting, rectification, component changes, NDT callouts and the certification time to close the tech log. Then calculate:
- Weekly mean: the background defect load.
- P80 week: the level you can cover in four weeks out of five.
- Worst credible week: the escalation case for management, not the normal roster target.
02Worked example: four AW139s
Take a four-aircraft AW139 fleet flying 2,400 hours a year. Your defect records show 960 unscheduled man-hours — 0.4 MH per flight hour. The weekly history looks like this:
| Planning input | Value | Planning meaning |
|---|---|---|
| Annual unscheduled hours | 960 MH | Actual rectification workload |
| Weekly average | 18.5 MH | 960 ÷ 52 |
| P80 weekly demand | 29.0 MH | Normal roster protection level |
| Worst observed week | 47.0 MH | Escalation / contractor trigger |
| Certifying-staff share | 35% | Release and independent checks |
| Productive hours per shift | 6.5 h | After briefings and admin |
If the committed check and line schedule already needs 132 MH this week, and the roster offers 168 productive MH, the apparent slack is 36 MH. That looks comfortable until you compare it with the distribution: the average defect week fits, the P80 week almost fits, and the worst week misses by 11 MH.
reserved shifts = defect buffer ÷ productive hours per shift
29 ÷ 6.5 = 4.5 engineer-shifts per week
So the practical roster instruction is not “keep someone flexible.” It is: ring-fence 4.5 shifts — one engineer for nearly the whole week — and make sure at least 1.6 of those shifts are certifying shifts (29 × 35% ÷ 6.5).
03Build the buffer by licence category
A buffer made of the wrong people is still a shortage. Split the historical hours into B1 and B2 work. On an AW139/AW189 fleet, I usually see mechanical defect volume dominate, but avionics troubleshooting has longer elapsed time and a smaller certifying pool. That means B2 coverage can fail at a lower man-hour total.
For each category, apply the same arithmetic:
- P80 B1 defect hours ÷ productive shift hours = reserved B1 shifts.
- P80 B2 defect hours ÷ productive shift hours = reserved B2 shifts.
- Compare reserved shifts with the roster after leave, training and check commitments — not with the establishment list.
If the calculation says 2.2 B2 shifts and you only have one B2 on days, the answer is not to average B1 slack across the gap. Cross-licensing, an on-call agreement, or pre-booked contractor cover is the fix.
04Protect the buffer from the check schedule
The buffer only works if it is visible. Put it in the weekly roster as a named line — “unscheduled defect reserve” — with the same status as an A-check input. When the hangar asks to borrow it for a check that is running late, the planner can make an explicit trade: borrow 6.5 hours from the reserve and accept that P80 protection drops to 22.5 MH.
This is also the cleanest trigger for buying help. If unscheduled demand exceeds the P80 reserve twice in a rolling month, or if scheduled work repeatedly consumes the reserve, you no longer have a flexible buffer — you have a structural shortfall. Re-run the numbers in the ManHourPlan calculator, then compare contractor hours with the cost of slipping the next input.
05Make it defensible at review time
Your AMC1 145.A.30(d) man-hour plan should show both sides of workload: scheduled maintenance from the forecast and unscheduled maintenance from history. Keep three numbers each month: actual defect hours, hours charged against the reserve, and reserve hours borrowed by scheduled work. Those figures turn “we were busy” into evidence.
The full availability-waterfall method is in the Part-145 man-hour planning guide. For an offshore base, cross-check the duty and rotation effects against the IOGP-690 §25C manning guide — a 12-hour shift is not 12 productive hours, and offshore defect response is where that assumption gets expensive.
Build the defect buffer from your own utilisation, headcount and licence split. ManHourPlan shows the weekly workload, reserved capacity and certifying-staff gap in one auditor-ready scenario.
Calculate your buffer →