Air France 447, some facts are now known

Posted on 04. Jun, 2009 by in Featured

Determining the cause of an aircraft lost at sea is a very difficult process; with deep water, currents and the distant locations, sometimes the cause cannot be definitively determined. My Memorial Day post is an example; we never found a single piece of Iron Claw 606, even after searching the area for days. Modern technology has helped in the process; ACARS (Aircraft Communication Addressing and Reporting System), satellite links, and modern radar tracking gives us a starting point.

Established facts:
-Air France 447 flew into a line of storms (ACARS message, Radar track).
-AF 447 began to experience flight control failures (ACARS codes sent automatically).
-The A 330-200 structurally failed in flight (ACARS codes and debris field).
-The A 330-200 has a history of flight control anomalies (ATSB).

The timeline:
*02:00Z AF 447 enters a line of weather.
*02:10Z ACARS message sent, auto pilot off and alternate law of fly by wire system engaged.
*02:11Z-02:13Z ACARS sends series of ADIRU/ISIS faults.
*02:13Z ACARS sends PIM1/SEC 1 faults.
*02:14Z ACARS sends vertical speed of cabin message.

First some clarifications; most International flights fly with an augmented crew. That means more than two pilots, thus one (except for takeoff and landing) is in a crew rest seat. When Air France officials said they did not know if the Captain was at the controls they meant that they did not know who was in crew rest and who was in the cockpit. All International crews are type rated in the aircraft they fly, that means technically they are a Captain. The actual position (Captain) is determined by seniority (date of hire) not training or experience. There are many rumors etc. circulating; I offer this as an example of their lack of applicability to the investigation.

Next some back ground; the A 330 series aircraft has documented flight control events resulting in erroneous inputs and in flight upsets.

“ATSB (Australian Transportation Safety Board) found that the A330’s primary ADIRU sent erroneous data (spikes) on many parameters to systems in the aircraft, including the primary flight computers, that resulted in the autopilot disconnecting and two violent pitch-down events. Disassembly of the units will not be done until EMI testing is complete in order to prevent disturbance to the hardware. After disassembly, individual modules will be tested separately.”

EMI is electromagnetic interference; the ATSB believes that EMI may have caused the anomaly in the ADIRU which then sent improper commands to the flight control causing “violent pitch-down events”. Take note of the sequence: Auto pilot off, violent maneuvers until the crew could disable the ADIRU.

Now let’s examine the timeline of the ACARS messages sent by AF 447: Auto pilot off and alternate law engaged, ADIRU/ISIS faults, PRI 1and SEC 1faults, finally vertical cabin speed fault message.

Alternate Law: If a minor failure exists (sensors disagreement, computer blackout, hydraulic system upset,etc) this law becomes the active one. Basically it provides the same control as the normal law, but without protections. Also, the roll control becomes “direct law-like” control.

Without protections means potential for violent movements, as noted by the ATSB in the previous A 330 event. ADIRU faults means the air data inertial reference unit was sending erroneous data to the primary flight control computer (PRI-1). In the case of Qantas flight 72, as noted by ATSB, that resulted in violent flight control input. In the case of QF-72 the nose pitched down. The AF-447 ACARS transmitions next noted a series of faults starting with PRIM-1 going through SEC-1. There are 5 flight computers 3 primary and 2 secondary. I do not know the sequencing in the case of failure however it shows that the fly by wire system was getting conflicting inputs and thus stepping through backups in an attempt to re-establish flight control integrity. The last ACARS message was the vertical speed rising in the cabin.

The ATSB believes that EMI was possibly responsible for the ADIRU spikes that caused the violent pitch event in QF-72. AF-447 was in a weather front where lightening or static discharge could have produced an EMI event. In the alternate law mode the flight control movements were unprotected, thus could have moved the primary flight control surfaces in or near full throw. If max Beta-Q (yaw dynamic pressure) was exceeded then the vertical stabilizer would have likely failed. Past Airbus models have had tail plane failures. The aircraft could have continued down range as it broke apart accounting for the 55 mile debris field.

In my opinion it is evident that flight control events preceeded the tragic in-flight break up of AF-447. With out the black boxes however it will be vary hard to come to an informed conclusion.

No Responses to “Air France 447, some facts are now known”

  1. Stephen Greaves 5 June 2009 at 05:25 #

    QANTAS is not spelt Quantas. It does not have a “u” because it is an abbreviation not a word. It stands for Queensland and Northern Territory Aerial Services.

  2. Chip 5 June 2009 at 12:13 #

    Yes I knew that, sometimes spell check fails you! Thanks for the correction I’ll fix it.

  3. Pedro Azevedo 6 June 2009 at 05:40 #


    It’s called a “PRIM”, not “PRI”.

    Otherwise, great job!

  4. Ernst Mulder 6 June 2009 at 07:53 #

    Could it be that the automated messages were sent _after_ a major structural failure of the plane?

  5. Chip 6 June 2009 at 22:44 #

    I think the time line is too long. For 3-4 minutes messages were sent. When the aircraft broke up IMO they would have stopped imediately due to interuption of electrical power and antenae seperation etc.

  6. Chip 6 June 2009 at 22:45 #

    I stand corrected, thanks.

Leave a Reply