Lap-joint corrosion beyond SRM limits on a lease-return narrow-body
During a 2C heavy maintenance check at a Hong Kong HKAR-145 organisation, low-frequency eddy-current inspection of a 13-year-old narrow-body's fuselage lap joints finds two areas of material loss on the same skin panel between stringers 14 and 16, frames 22 to 24: one 260 mm long with 0.55 mm loss, and one 90 mm long with 0.92 mm loss. The skin is 1.6 mm nominal 2024-T3 clad. The SRM permits blend-out to 0.5 mm loss followed by repetitive inspection; anything beyond that requires repair. The airframe is in the hangar with the cabin floor open, the operator has an AOG on the next rotation in six days, and the aircraft is due for lease return in nine weeks with redelivery tolerances covering repairs. A colleague suggests blending the deeper area 'a bit further so it looks uniform and measures inside limits'.
1. Assess whether either area can be released under the SRM alone. Set out the decision logic and the evidence you would require before accepting any measurement.
Read the worked answer
Conclusion first: neither area can be released on the SRM. The logic is a straight comparison against an approved limit, but the limit only means something if the measurement is valid and the allowance actually covers this configuration.
Step 1 - Validate the data. Cladding makes eddy-current readings sensitive to lift-off and surface condition; I would want the NDT procedure reference, probe and calibration records, inspector certification, surface preparation, a repeat scan, and an independent second inspection by another Level 2. If the 0.55 mm figure is within probe tolerance of 0.5 mm, it is still not a pass - it is a borderline result needing engineering review, not a rounding-down exercise.
Step 2 - Confirm applicability. Check that the SRM allowance is for clad 2024-T3 lap joints, that it addresses the joint type, fastener rows, frame and stringer proximity, and whether the repetitive inspection route is available on this panel. A blend-out allowance is not a repair approval for a different damage mode such as intergranular attack or exfoliation underneath.
Step 3 - Reject the suggestion outright. Blending removes material. Blending to make a 0.92 mm reading 'look uniform' removes further thickness from a fatigue-critical lap joint, destroys the evidence trail, and misrepresents the airworthiness status on a certificate of release. That is a data-integrity violation, not engineering judgement. The 0.92 mm area is well outside the allowance and needs an approved repair.
Step 4 - Decide the path. Hold the panel, photograph and map the damage, and raise both areas with the type certificate holder or an approved design organisation for disposition.
2. The operator wants the aircraft back in six days. Describe the approval route for a repair, the substantiation you would demand, and how you would sequence the work.
Read the worked answer
The repair is almost certainly major, so the organisation performing it cannot approve its own design. The route is an approved design under HKAR-21 from the type certificate holder or a design organisation with repair approval authority; the continuing airworthiness manager classifies minor versus major and the design data must come from the approved source. Blending deeper without that data is not an option.
Substantiation I would require:
- Static strength of the repaired net section, including load redistribution to adjacent frames and stringers.
- Fatigue and damage tolerance: lap joints are susceptible to widespread fatigue damage, so the repair must show either continued inspection or life limitation, plus its effect on the existing inspection programme.
- Corrosion control: surface treatment, sealant system, fastener type and interference fit, and material compatibility with clad skin.
- A defined inspection method and repeat interval after repair, and a redelivery position acceptable to the lessor.
The risk here is schedule optics. A six-day AOG target is achievable only if the design request goes out on day zero with a complete data pack: NDT maps, skin thickness survey, material and clad condition, photos, fastener layout and aircraft hours and cycles. I would keep the floor open, prepare the work package and tooling in parallel, and not close up until the approved data is in hand.
A permit to fly for the transit to a repair base is not a shortcut; it requires an approved justification that the aircraft can be flown safely, and unrepaired damage beyond the SRM cannot be waved through by a commercial schedule.
3. What records and continuing airworthiness actions follow the repair, and what would you tell the lease-return team?
Read the worked answer
Records. The certificate of release must reference the approved repair data explicitly, not a generic statement. The package should contain the design approval reference, drawings, the NDT report with the pre-repair damage map, material and fastener certificates, the post-repair inspection results, and the torque and sealant records. Blend-outs within the SRM also need recording with dimensions so the next check is measured against the right baseline.
Continuing airworthiness. The maintenance programme must be amended to capture any new or changed inspection, and the repair must carry an airworthiness limitation if the design data imposes one. Check whether the repair interacts with any existing airworthiness directive or supplemental structural inspection programme. Re-run the weighing and balance check if the repair adds meaningful mass, and update the aircraft statement of conformity accordingly.
For the lease-return team, the honest message is that the aircraft no longer conforms to the original build standard at that panel, and the lessor's technical acceptance will depend on the acceptability of the approved repair and its documentation. Redelivery tolerances usually allow approved repairs but require disclosure, so I would send the design approval reference and damage map early rather than let the lessor discover it in the records review. Do not let a commercial redelivery deadline drive a decision to blend outside limits: a repair that is disclosed and approved is a schedule problem, whereas an undisclosed one is a certificate and insurance problem.