A little more
1994 | 1864 | Flight Archive
A330 crash caused by series of small errors
DAVID LEAR MOUNT
A series of small errors combined to cause the crash of the Airbus A3 30 during tests at Toulouse on 30 June, says the official inquiry's interim report.
Describing the events that took place at Blagnac airfield, the report portrays a tired captain, flying at the end of an extremely busy day with an incompletely briefed crew.
The commission of enquiry within the DGA, the French military body which investigates test flight accidents, highlights "...a combination of several factors, no one of which, in isolation, would have caused the crash".
The fatal take-off was the second in a sortie designed to test the new autopilot system under critical engine-out conditions immediately after take-off.
During the first take-off, the crew had carried out the test successfully, then flew engine-out goarounds twice. The crew then made a full-stop landing and taxied- round for another take-off, by which time the sortie had already lasted for 55min.
The captain, Airbus chief test pilot Nick Warner, had flown the first take-off, but for the second he gave control to the co-pilot, Michel Cais, an Air Inter training captain who had been working with the Airbus training organization Aeroformation.
Warner must have been tired after a punishingly busy day, says the report. Before starting the test flight, he had already captained an A321 demonstration flight, supervised a simulator session and attended two meetings, including a briefing for journalists.
Criticising the lack of a complete pre-flight briefing, the report relates this to Warner's busy schedule. It also remarks that the crew may have become complacent because the tests up to "the last take-off" had gone perfectly.
Under the heading "probable causes", the Commission then lists three groups of contributory factors, the first of which were related to the test-flight tasks immediately after take-off.
Maximum thrust [take-take-off/goaround (TOGA)] was chosen by the captain, instead of setting "Flex 49", which was die high power setting specified in the test plan. This meant that, when he simulated failure of the left engine immediately after take-off, asymmetric forces were higher than planned.
The aircraft centre of gravity (CG) was at maximum aft setting of 42%, while the trim was set at 2.2° nose-up. Although the trim setting was within acceptable limits, it was inappropriate, given the aft CG. These two factors together increased the risk of fast over-rotation.
The autopilot was also left set for altitude-capture at 2,000ft (600m), which meant that the mode became active within seconds of take-off.
The report warns that the absence of attitude-protection in the autopilot altitude-acquire mode "played a critical role" in the accident.
According to the report, die next group of factors related to uncertainty in the allocation of tasks between captain and co-pilot. The co-pilot rotated the aircraft for take-off "firmly and very fast" to an attitude more than 25°, compared with 14.5° at first take-off.
Immediately after take-off, the captain then carried out the test procedures: autopilot engage, throttle-back port engine and trip circuit-breaker for blue hydraulic circuit. The report says that he therefore became temporarily "out of the piloting loop".
Ancillary factors played an equally important part in the accident, comments the report.
These include:
• the lack of autopilot-mode visual indication because the exceptional pitch attitude obscured it on the primary flight display;
• crew confidence in how they expected the aircraft to react;
• the delayed reaction of the test engineer to developing parameters, particularly speed;
• "the captain's slowness in reacting to the development of an abnormal situation".
Interim recommendations
• Advise A3 30 operators of the need to monitor aircraft airspeed behaviour if they are in autopilot altitude-capture mode during the climb and an engine fails;
• conduct tests to ensure that this mode poses no risks in line usage;
• whatever is found, seek solutions to improve flight-envelope protection in this mode to make it the same as in all other modes;
• carry out similar tests on all other aircraft types which have similar autopilot systems;
• review the regulations concerning who, in addition to the basic crew, may fly on test flights;
• test flights should be preceded by a formal briefing, even if the test is routine. This should include any planned, authorised re-allocation of tasks during the flight.