Accident 24 July 2014 in Gossi, Mali to the MD-83 registered EC-LTV operated by Swiftair S.A.
Press release from the BEA and the Commission d’Enquêtes Accidents et Incidents de l’Aviation civile (Mali)
Following the publication of the Interim Report on 20 September 2014 in Bamako (Mali), investigative work has continued, based on the analysis of the accident flight parameters. Progress made in this work has led the Republic of Mali Commission of Inquiry and the BEA to communicate jointly the following information.
On 24 July 2014, the MD-83 registered EC-LTV was performing scheduled night flight AH 5017 from Ouagadougou (Burkina Faso) to Algiers (Algeria). Takeoff occurred at 01h15, the climb towards the cruise altitude took place without any significant events, and the crew made several heading changes in order to fly around a storm cell. The autopilot and the autothrottle were engaged. The aeroplane reached the cruise altitude of 31,000 ft, that’s to say about 9,500 m. The autopilot then switched to the mode that maintains the altitude and the autothrottle to the mode that maintains the speed (Mach).
About two minutes after levelling off at an altitude of 31,000 ft, calculations performed by the manufacturer and validated by the investigation team indicate that the recorded EPR , the main parameter for engine power management, became erroneous on the right engine and then about 55 seconds later on the left engine. This was likely due to icing of the pressure sensors located on the engine nose cones. If the engine anti-ice protection system is activated, these pressure sensors are heated by hot air.
Analysis of the available data indicates that the crew likely did not activate the system during climb and cruise.
As a result of the icing of the pressure sensors, the erroneous information transmitted to the autothrottle meant that the latter limited the thrust delivered by the engines. Under these conditions, the thrust was insufficient to maintain cruise speed and the aeroplane slowed down. The autopilot then commanded an increase in the aeroplane’s pitch attitude in order to maintain the altitude in spite of this loss of speed.
This explains how, from the beginning of the error in measuring the EPR values, the aeroplane’s speed dropped from 290 kt to 200 kt in about 5 minutes and 35 seconds and the angle of attack increased until the aeroplane stalled.
About 20 seconds after the beginning of the aeroplane stall, the autopilot was disengaged. The aeroplane rolled suddenly to the left until it reached a bank angle of 140°, and a nose-down pitch
The recorded parameters indicate that there were no stall recovery manœuvres by
However, in the moments following the aeroplane stall, the flight control surfaces remained deflected nose-up and in a right roll.
At least two similar events occurred, in June 2002 and in June 2014, with no serious consequences.
The event in June 2002 was the subject of an NTSB investigation report. On 4 June 2002, the McDonnell Douglas MD-82, registered N823NK performing Spirit Airlines flight 970, suffered a loss of thrust on both engines, in cruise at an altitude of 33,000 ft, that is about 10,000 m. The two pressure sensors, located on the engine nose bullets, were blocked by ice crystals, leading to incorrect indications and over-estimation of the EPR. The crew noticed the drop in speed and the precursor indications of a stall just before disengagement of the autopilot and putting the aeroplane into a descent. They had not activated the engine anti-ice systems. This event occurred during the day, outside the clouds.
On 8 June 2014, the MD83 registered EC-JUG belonging to Swiftair, which was performing a passenger transport flight at flight level FL 330, suffered a drop in speed while it was flying during the daytime above the cloud layer. The crew detected the problem, put the aeroplane into a descent and activated the engine anti-ice systems without reaching a stall situation, then continued the flight.
This background, as well as the data on the accident to flight AH5017, was shared with the European Aviation Safety Agency (EASA) and through EASA with the American authorities (FAA); they should serve as the basis for future publication of corrective measures aimed at assisting crews in identifying and responding to similar situations to those encountered at the time of this accident.
The investigative work is continuing, in particular on the analysis of:
the flight parameters to complete the scenario described above,
possible crew reactions, despite the absence of Cockpit Voice Recorder data from the accident flight, which remain unusable to this day,
the training and follow-up of Swiftair crews,
previous events and the follow-up undertaken.
The publication of the final report is planned before the end of December 2015.
Download Interim below: