What happened
On July 22, 2011, an Airbus A340-313, registration F-GLZU, operated by Air France, was cruising at FL350 over the North Atlantic near 18N 060W. While flying through moderate turbulence, the aircraft encountered a headwind gust that triggered an "OVERSPEED" warning.
In response to the sudden alarm, the pilot monitoring (PNF) manually disconnected the autopilot. This action was accompanied by an instinctive, involuntary nose-up command on the sidestick lasting approximately six seconds. Simultaneously, the pilot flying (PF) deployed speed brakes. The combination of the upward gust, the manual pitch input, and the deployment of speed brakes caused the aircraft to climb rapidly, reaching a peak altitude of 38,150 ft with a vertical speed of 5,700 ft/min. During this excursion, the aircraft's airspeed dropped to 0.66 Mach. The crew eventually regained control, re-engaged the autopilot, and stabilized the aircraft at the assigned FL350.
The investigation
The BEA investigation examined flight data from the FDR and maintenance recorder, as well as satellite imagery and meteorological data. The investigation focused on the sequence of cockpit alerts, the aircraft's longitudinal behavior, and the crew's monitoring of flight parameters.
Investigators analyzed the effectiveness of the aircraft's high-angle-of-attack protections and the visibility of cockpit alerts. A significant challenge for the investigation was the lack of Cockpit Voice Recorder (CVR) data, as the incident occurred eight hours before landing, exceeding the 2-hour recording capacity of the installed unit. The investigation also reviewed the crew's use of the weather radar and the settings used for the navigation display (ND) scale.
Findings
- The primary cause of the incident was inadequate monitoring of flight parameters, which resulted in the crew failing to detect the autopilot disconnection and the subsequent altitude deviation.
- The pilot monitoring experienced a "startle effect" due to the unexpected overspeed alarm, leading to an instinctive and inappropriate manual input.
- The autopilot disconnection alarm was not effectively perceived because the higher-priority "OVERSPEED" alarm masked the auditory signal.
- The crew's attention was diverted by secondary tasks, such as managing meal trays and making passenger announcements, during the period of altitude deviation.
- The weather radar settings were not optimal for detecting the convective cells encountered, as the ND scale was set to 320 NM rather than the recommended 160 NM or 80 NM.
- Turbulence made the visual reading of flight instruments more difficult.