Avionics News March 2012 - 64

AIR FRANCE FLIGHT 447
Continued from page 63

Captain. The airplane went into a stall that lasted until the impact with the sea,” which occurred at 0214:28. When the autopilot disengaged, followed a few seconds later by the autothrottle system, the A330’s two co-pilots noted loss of airspeed indications on their primary flight displays and the integrated standby instrument system, an all-in-one attitude/airspeed indicator and altimeter. Nose-up control inputs resulted in the stall warning sounding twice and the airplane’s pitch attitude increasing to 10 degrees nose-up; the A330 began to climb, eventually at 7,000 feet per minute. The stall warning sounded again. Thrust levers were placed in the takeoff/go-around position, and noseup pitch commands were maintained. The A330 climbed to about 38,000 feet (flight level 380) before beginning a descent. By the time the captain returned to the cockpit from the rest area, the stall warning had stopped. The airplane’s angle of attack exceeded 40 degrees nose-up, and it was descending at approximately 10,000 feet per minute. Roll oscillations reached 40 degrees. Shortly, pitch-down control inputs were made, which reduced the angle of attack. The stall warning sounded again, and airspeeds became valid. The crew once again applied nose-up pitch inputs; the stall warning ceased, and airspeed values dropped. The airplane hit the water in a descent of 10,912 feet per minute and a groundspeed of 107 knots. Its attitude was 16.2 degrees nose-up, with a 5.3-degree bank to the left on a magnetic heading of 270 degrees. The BEA’s third interim report on AFR447, which summarized the agency’s findings as of July 29, 2011, concluded the blocked pitot tubes caused the A330’s flight control system to note erroneous airspeed data was being received and disengage the autopilot and autothrottle systems, as it was designed. All of a sudden, the crew members – who had been mostly worried about circumnavigating a thunderstorm – were asked to manage the tasks of maintaining altitude, speed and heading. They had years of experience but couldn’t pull it off. Why? As with any significant mishap involving a modern jet transport, the answers are complicated. Although its final report had not been published as this article was being prepared, the BEA already has made recommendations involving flightcrew training in “specific and regular exercises dedicated to manual aircraft handling of approach to stall and stall recovery, including at high altitude” and “additional criteria” for relief captains. Other recommendations include installing angle-of-attack instrumentation and using cockpit
64
avionics news

video recordings to supplement the flight data and cockpit voice devices. Given the emerging picture of what happened, these early recommendations are both appropriate and expected. Additional ones are anticipated as part of a final report. But, for many observers, what the AFR447 crash highlights is how far removed from the human/machine loop the flightcrew allowed themselves to be drawn. How can that be addressed, and what lessons can the avionics industry take away?

TURNING POINT
The BEA’s interim reports make it clear the flightcrew had little to no idea what was wrong with their jet, what it was doing or how to recover. Basically, two qualified pilots watched for more than three minutes as a third maintained nose-up control inputs, placing the airplane in an aerodynamic stall and 38,000foot descent from which it did not recover. Unless a subsequent report sheds additional, unforeseen light on systems- or instrumentation-related reasons for this outcome, one of the BEA’s official findings certainly will include indictments of the pilots and their training. At the same time, this tragedy may sadly present opportunities for avionics manufacturers. Already, the BEA has recommended installing angle-of-attack instrumentation in cockpits. Others have looked at the remoteness of the crash site and difficulties in even realizing when and where the airplane had gone down to propose new devices and technologies. Among them are emergency locator transmitters, which eject themselves from the airplane in a crash. Another potential technology is real-time transmission of CVR/FDR information to ground stations or via satellite, so the data investigators need isn’t hidden under three miles of water. The extremely limited ACARS data AFR447 transmitted in its final minutes gave investigators tantalizing clues of what happened, but little real progress was made until the all-important flight recorders were retrieved and analyzed. Bandwidth and data transmission expenses, along with flightcrew privacy concerns, make a real-time CVR/FDR solution problematic, however. The good news is the “traditional” CVR/FDR devices worked as advertised, and perhaps well beyond its design parameters. Meanwhile, much has been said and written during the last couple of decades debating the impact of highly automated cockpits in aircraft ranging from two-seat trainers to long-haul transports, such as the Airbus A330. For the avionics industry, the AFR447 crash highlights how expectations of accuracy, reliability and repeatability in cockpit automation have succeeded, but at the cost of reducing pilots’ ability to remain in the human/ machine loop. Pilots often joke about their understanding of the automation with phrases such as, “Now what’s it doing?” or “Yeah, it does that.”

•

march

2012

Avionics News March 2012

Table of Contents for the Digital Edition of Avionics News March 2012