The US Federal Aviation Administration looks as if it might be getting the message about the need for airline pilots to undergo upset recovery training.

It has taken many fatal loss of control (LOC) accidents to concentrate the FAA's mind. But despite the FAA's example and the fact that safety analysts have long noted an increase in the relative frequency of such accidents, there is no sign that the European Aviation Safety Agency is suffering a similar attack of common sense.

This subject, aired on's Learmount blog, has attracted informed comment about upset recovery training, especially since LOC at night may have been one of the factors in the crash of Air France AF447. A litany of LOC accidents, including recent ones like Adam Air and Flash Airlines, is listed there, whereas in its current argument the FAA cites US upsets - specifically the Pinnacle Airlines Bombardier CRJ200 and Colgan Air Q400 crashes.

The FAA proposal is about how upset recovery training could be carried out effectively, not whether it is required. The agency has in mind to mandate training of a type that's available, but is not much used because it is expensive. The FAA starts from the assumption that simulators are inadequate because they cannot provide the acceleration feedback, and there is insufficient performance data for, say, a Boeing 737 at 130° bank with 50° nose up, to enable a simulator to replicate reliably the results of pilot input. So the FAA may mandate airborne training in a limited aerobatic, variable stability Learjet. Sure, that would be desirable, because it ticks all the boxes, including providing something akin to the inertial reactions of an airliner.

But light aeroplane training is much less expensive. Maybe the FAA is worried pilots would over-react to upset recovery training as happened in the 2001 American Airlines AA587 wake encounter, but even training in the Learjet would provide no guarantee against that.

Assuming no external visual cues, upset recovery training has to teach the pilot to mentally reject his/her sensory impressions, and react only to what the flight instruments reveal. Human sensory and balance organs are easily fooled. Teaching this hard discipline is critical, and an ideal training programme would include airborne practice.

But Boeing has discovered (see blog) that well-designed simulator training that concentrates purely on reacting correctly to the instrument picture is better than nothing.


Source: Flight International