ATSB Preliminary Factual Report, In-flight upset, Qantas Airbus A330, 154km west of Learmonth, WA, 7 October 2008
The following is the text of the media briefing given by Ms Kerryn Macaulay, Director, Strategy and Capability, at 10am 14 November 2008.
Good morning. I am releasing the Australian Transport Safety Bureau's Preliminary Factual report on the in-flight upset of an Airbus A330 aircraft, registered as VH-QPA, which occurred 154 km west of Learmonth, Western Australia, while being operated on a scheduled passenger flight (Qantas Flight 72) from Singapore to Perth on 7 October 2008. The ATSB immediately commenced an investigation into this accident[i] and has been working collaboratively with our safety investigation counterparts in France and the United States - the French BEA (Bureau d'Enqutes et d'Analyses pour la s'curit' de l'aviation civile) and the US NTSB (National Transportation Safety Board), as well as the aircraft and component manufacturers, Qantas, and CASA.
As you know, at about 1240 on 7 October 2008, while the aircraft was cruising at 37,000 ft, the autopilot disconnected, accompanied by various aircraft system failure indications. While the crew was evaluating the situation, the aircraft abruptly pitched nose-down and descended 650 ft. After returning the aircraft to 37,000 ft, the crew commenced actions to deal with multiple failure messages. Shortly after, the aircraft commenced a second uncommanded pitch-down event and descended about 400 ft. The crew's timely response led to the recovery of the aircraft descent within seconds in relation to both pitch-down events.
The crew initially made a PAN emergency broadcast to air traffic control but upgraded this to a MAYDAY after receiving advice from cabin crew on the extent and seriousness of injuries incurred by passengers and crew. The aircraft diverted to land at Learmonth and this was done carefully by the flight crew as is detailed in the ATSB report.
Unfortunately, during the upset, a flight attendant and at least 13 passengers were seriously injured and many others experienced less serious injuries. Most of the injuries involved passengers who were seated without their seatbelts fastened.
Examination of flight data recorder information indicates that, at the time the autopilot disconnected, there was a fault in a flight computer system component known as the air data inertial reference unit number 1 (ADIRU 1)[ii] which resulted in a number of spurious spikes in ADIRU parameter values. Further spurious parameter spikes continued to influence a number of system failure indications throughout the flight, resulting in frequent failure messages being provided to the crew. The crew completed required actions in response to the messages, but these actions were not effective in removing the spikes or failure indications. The investigation team is continuing to examine the influence of the spikes in ADIRU parameters on the performance of the flight controls[iii].
Most components on modern aircraft, including ADIRUs, are highly reliable and there has only been a small number of occasions where ADIRUs of different types made by varying manufacturers have had some form of failure. It is extremely rare for any such failures to have an effect on an aircraft's flight controls. The ATSB has previously investigated an in-flight upset related to ADIRU failure from a different manufacturer in a Boeing 777 which occurred in 2005 and was traced to a software fault. While a software fault has not been ruled out in the current investigation, it seems unlikely that the two events are linked.
In terms of ongoing investigation activities, the ATSB is able to advise the following:
- The three ADIRUs will be subject to comprehensive testing at the manufacturer's facilities[iv] in the US. This testing is planned to commence on Monday. A carefully prepared test plan is currently being finalised in anticipation of this complex work to ensure the investigation team has the best possible chance to understand what led to the pitch-down events in order to provide a basis to eliminate the problem at its source.
- The investigation will review the ADIRUs' data monitoring capability and management of anomolous ADIRU data, including flight deck indications and will also review records of previous occurrences involving ADIRU failures (which did not result in in-flight upsets) and any occurrences where large numbers of spurious messages were generated.
- Subject to the results of the ADIRU testing, examination of other aircraft components may be conducted such as the three flight control primary computers and their software in order to understand why the fault in the ADIRU was able to be translated to flight control movements.
- Possible external sources of electromagnetic interference are being explored and assessed, including from the Harold E. Holt very low frequency transmitter near Exmouth, WA and from portable electronic devices on board the aircraft. This is unlikely, especially if the problem is clearly identified during the ADIRU and system testing.
- Work is still continuing on the cabin safety issues with interviews having been conducted with all the cabin crew as well as some of the passengers who were seriously injured. On 28 October 2008, the ATSB started distributing a passenger questionnaire seeking passenger observations during the upset events and asking questions in relation to the use of seatbelts, injuries and the use of personal electronic devices. Contact details for some passengers are incomplete. If any passenger has not received a questionnaire, please contact the ATSB on 1800 020 616 (or 61 2 6257 4150 from outside Australia) or via email to email@example.com. A review of relevant industry requirements regarding the use of seatbelts is also being conducted.
- The ATSB is aware that a post-incident multi-agency debrief of the emergency response to the accident has been conducted in Western Australia. The ATSB will review the outcomes of that debrief in relation to information obtained at interviews and from responses to the passenger questionnaire.
A number of important safety actions have already been implemented arising from the investigation to date. These include:
- Following a 14 October 2008 telex, Airbus issued an Operations Engineering Bulletin on 15 October, applicable to all A330 aircraft fitted with Northrop-Grumman ADIRUs which detailed a procedure for flight crew to follow in the event that specified fault indications were observed in order to reduce or eliminate the risk of a future similar in-flight upset event.
- The aircraft operator has incorporated the material from Airbus in a Flight Standing Order for its A330 operations and has commenced a program of focussed training during simulator sessions and route checks to ensure that flight crew undertaking recurrent or endorsement training are aware of the contents of the Flight Standing Order.
- On 27 October 2008, the Civil Aviation Safety Authority issued a media release to reinforce the ATSB's message that the occurrence was a timely reminder to 'remain buckled up when seated at all stages of the flight'.
The investigation is ongoing and the ATSB continues to work closely with the BEA, the NTSB, Airbus, Qantas, Northrop Grumman and CASA. It is always difficult to predict how long an investigation such as this will take. While it is likely to take some time, should any critical safety issues emerge that require urgent attention, the ATSB will immediately bring such issues to the attention of the relevant authorities who are best placed to take prompt action to address those issues.
[i] As serious injuries were incurred, this constituted an accident under the International Civil Aviation Organisation definition outlined in Annex 13 to the Chicago Convention and as defined in Australia's Transport Safety Investigation Act 2003.
[ii] There are three ADIRU units that form part of the Air Data and Inertial Reference System.
[iii] There are three flight control primary computers (commonly known as PRIMs) in the aircraft's flight control system with one PRIM operating as the master in normal operations. It processes and sends orders to other computers which may result in flight control surfaces being moved to execute different manouevres including to pitch the aircraft's nose up or down. One source of information to the PRIMs are the ADIRUs.
[iv] Northrop Grumman Corporation.Media contact: 1800 020 616 Last update 01 April 2011