Aloha Airlines Flight 243

A routine interisland flight

Aloha Airlines flight 243 departed Hilo for Honolulu at 13:25 Hawaiian standard time on 28 April 1988. Boeing 737 N73711 had already flown interisland sectors that day without a reported system or structural abnormality. The departure and climb were uneventful, and the weather was not a factor in the accident. The flight carried 89 passengers, two pilots and three flight attendants, with an FAA air traffic controller occupying the cockpit observer seat. That observer explains why the executive summary groups six people with the crew while the injury table separates one as an other occupant.

As the aircraft levelled at 24,000 feet, the pilots heard a loud sound followed by wind noise. Debris entered the cockpit, the cockpit door was missing, and the captain could see open sky where the forward cabin ceiling had been. At 13:46 a large portion of upper fuselage skin and structure separated, producing explosive decompression. About eighteen feet of cabin structure behind the entrance door was lost. Cabin air expanded and escaped through the opening.

One flight attendant was swept from the cabin and was presumed fatally injured after an unsuccessful sea search. Another was struck by debris and seriously hurt. Passengers remained seated with the seat-belt sign illuminated at the initiating event. The surviving cabin crew continued to assist and reassure them despite injuries and the exposed cabin. The crew had to fly and land an aircraft whose structure, systems and cabin environment had been severely altered.

Descending and landing with several emergencies

The captain took over the controls and began an emergency descent. Both pilots and the observer donned oxygen masks. Wind noise initially prevented normal conversation, so the pilots used hand signals. The first officer selected the emergency transponder code and tried to alert air traffic control before establishing contact with Maui Tower. Kahului on Maui was the diversion landing airport. The communication difficulties were a direct operational consequence of the missing cabin structure, adding to the task of controlling the damaged aircraft and assessing what remained functional.

During the approach the captain found the aircraft less controllable when flaps were extended beyond five, and returned them to that setting. He also chose an approach speed of 170 knots because control deteriorated below it. The main landing gear indicated down, but the nose-gear down indication did not illuminate even after manual-extension attempts. A direct visual check was not made because of the occupied centre jumpseat and the urgency of landing. The left engine also failed, and an attempted restart did not restore it.

Flight 243 landed on Kahului's runway 02 at 13:58:45. The captain reported a normal touchdown and rollout, using the remaining engine's reverser and the brakes. An emergency evacuation followed on the runway. Eight people had serious injuries, including seven passengers and one flight attendant. The aircraft was damaged beyond repair.

The first officer tried to reach the flight attendants through both the public-address system and the interphone, without a response. She told the tower that the crew could not communicate with them and would need help for the passengers after landing. That call did not result in an ambulance request. She supplied the passenger count, but did not give the fuel load. The controller did not repeat the request for fuel information even after the emergency-response chief asked about it. These gaps affected preparation on the ground while the pilots concentrated on controlling the aircraft and arranging an immediate landing.

Reconstructing a missing piece of fuselage

The initial failure origin was not recovered. NTSB therefore examined the structure still attached to the aircraft and its service history to reconstruct where separation began. Deformation, broken frames and floor damage pointed to the left side of the upper fuselage. The Board concluded that failure initiated at the lap joint along S-10L. The final findings describe multiple-site fatigue cracking near rivet holes, together with tear-strap disbonding, as the mechanism that defeated the intended fail-safe behaviour.

The early fuselage lap joints depended partly on bonded surfaces to transfer loads. Manufacturing difficulties in surface preparation or bond processing reduced environmental durability or left portions insufficiently bonded. When bonding deteriorated, loads transferred through the rivets and concentrated at the knife edges associated with countersunk holes. Fatigue cracks could then develop at many sites. Numerous small cracks could link into a larger opening, while disbonded tear straps no longer reliably contained the damage.

The intended structural response was a limited opening and controlled decompression, with the surrounding structure retaining adequate integrity. NTSB found that sufficient cracking and disbonding existed to negate that response. After the accident, a passenger reported seeing a longitudinal crack while boarding at Hilo but had not told airline staff or the pilots.

A programme that did not match the aircraft use

N73711 had accumulated 89,680 flight cycles and 35,496 flight hours. Short interisland services generated many cycles in relatively little flying time. Some short flights reached only partial pressure differential and imposed partial loads on the fuselage. Even after allowing for those flights, cycle accumulation exceeded the maintenance-planning assumptions. Inspection intervals built primarily around flight hours could allow substantially more cyclic exposure than the original maintenance-planning assumptions anticipated.

NTSB identified high cycle accumulation between structural inspections, long elapsed intervals for related deterioration and a heavily segmented inspection programme. Inspecting small portions over an extended period limited the ability to assess the aircraft's overall structural condition and carry out full repairs. Overnight scheduling and the need to return aircraft to service constrained time and manpower. The scheduling constraints limited both inspection and repair work.

The Board found that Aloha had sufficient information about lap-joint problems to implement an effective detection and repair programme. That information included known bonding and corrosion issues and prior damage in the fleet. It also found uncertainty about a required eddy-current inspection: the investigation could not determine whether it had not been performed or had been ineffective.

Investigators compared fatigue striations in surviving cracks with the service accumulated since the previous inspection. This allowed them to estimate the size of cracks at the time that inspection was due. Postaccident eddy-current checks by Aloha inspectors reliably detected cracks above the demonstrated threshold. Some cracks were estimated to have exceeded that detectable size at the earlier inspection. The NTSB therefore concluded that the required process had either been omitted or had failed to find damage that the method should have detected. The surviving structure supplied evidence for that conclusion even though the initial failure origin had been lost.

Inspection, oversight and design assumptions

Human performance was a substantial part of the inspection problem. NTSB examined the demands of repetitive visual and nondestructive tasks, the working environment and the effects of maintenance scheduling. Repetitive rivet inspections imposed demands on attention, access and time, with fatigue affecting the reliability of the work. The Board criticised management's failure to recognise these factors and focus inspectors on the critical nature of crack detection, corrosion control and lap-joint inspection. These are organisational conditions that affect the reliability of the technical method.

FAA oversight also failed to establish the actual condition of the operating fleet. The report describes a surveillance approach focused too heavily on manuals and records, and an assigned inspector burdened with other responsibilities and insufficiently informed about the fleet and its operations. Required lap-joint inspection had also been limited more narrowly than Boeing's proposed scope. The Board included FAA programme evaluation and inspection requirements among the contributing factors.

The design and certification analysis considered why the deterioration escaped the original controlled-decompression assumptions. Fatigue tests had not reproduced the long-term effects of bonding degradation, corrosion and multiple-site cracks found in service. Supplemental inspection classifications that treated fuselage skin damage as readily obvious were also questioned. NTSB's final cause attributed the accident to the maintenance programme's failure to detect significant disbonding and fatigue, with management, FAA oversight and the absence of complete corrective action for early cold-bond problems contributing.

The Board examined the physical demands of the rivet inspections. Inspectors had to work from scaffolding, move along the upper fuselage and carry a bright light. Eddy-current inspection required a probe and meter as well. Work on top of the fuselage sometimes required ropes attached to the hangar rafters to prevent a fall. Repeating close inspections over long rows of rivets created demands on attention and expectation. The analysis identified boredom, task length, isolation and the surrounding environment as influences on reliable performance. Night work added concerns about sleep loss, irregular rest schedules and circadian effects on mechanics and inspectors.

Safety action and crew performance

Recommendations addressed the maintenance programme itself, inspector competence and the evaluation of real aircraft condition. NTSB sought engineering support for approval of segmented inspections, programme limits suited to cycle and time exposure, modern maintenance curricula and recurrent training for nondestructive inspection personnel. It also called for corrosion control programs and evaluation of maintenance effectiveness using engineering resources. The Board sought programmes capable of finding and correcting progressive structural deterioration.

The Board commended the flight and cabin crews for managing an unprecedented emergency and bringing the aircraft to a safe landing. Its analysis also identified a concern with the emergency descent: using high speed and speed brakes without first adequately accounting for structural integrity could impose additional loads on the damaged aircraft and expose passengers to further harm.

The training recommendation included apprenticeship and periodic demonstration of skills for inspectors performing nondestructive work. The Board also wanted modern industry technology included in mechanic-school curricula and licensing tests. Programme approval required engineering support where airlines proposed dividing major inspections into smaller segments. Another recommendation asked the FAA to identify operators whose relationship between flight cycles and flight time differed substantially from maintenance-planning assumptions. Their programmes would then need to provide timely detection of both cycle-related and time-related deterioration. These proposals addressed the frequent short sectors flown by Aloha and the practical reliability of the inspections intended to support continued service.