 |
| Credit; CAE |
There is some concern, though, that the actual costs may be higher when airlines attempt to obtain from airframers or third-party providers the “extended” aerodynamic envelope models needed for simulators to accurately reflect the handling characteristics of their fleets.
We won’t have to wait long to find out—carriers will need those models before March 2019 to begin providing pilots with specific simulator training in aerodynamic stalls, upset prevention and recovery, airframe and engine icing, bounced-landing recoveries and takeoffs and landings in gusty crosswinds. The five training enhancements were mandated by Congress in the 2010 Airline Safety FAA Extension Act, which was driven by the February 2009 loss-of-control (LOC) crash of Continental Connection/Colgan AirFlight 3407 and a slew of NTSB recommendations.
To accommodate the upgrade, a portion of the existing cadre of approximately 335 flight simulation training devices U.S. airlines use for training will have to be assessed and evaluated by subject-matter-expert (SME) pilots and most likely upgraded to handle the new training areas.
A key part of the equation was delivered on March 30, when the FAA issued a final rule defining how simulators must perform in the “extended” envelopes needed for the training, some of which are outside the normal pitch, roll, yaw and angle-of-attack (AOA) bounds.
Legacy full-motion simulators, which typically cost $8 million or more, are very representative of actual aircraft within certain bounds. For aerodynamic stalls—a factor in the Colgan and many other LOC crashes—the devices since 1980 have had to accurately model flight maneuvers up to and including a full stall (an aerodynamic stall occurs at the stall AOA), with pilots trained to recover at the “first indication” of a stall. First indication on some aircraft might be a “buffet” action on the wing, particularly for fly-by-wire aircraft when the envelope protections have failed, and on others, a stick pusher activating.
However, accidents such as Colgan, Air France Flight 447 and the recent AirAsia Flight 8501 have shown that pilots in some cases find or put themselves well past the stall AOA, a realm they most likely have never seen and do not know how to recover from. Training for upsets and full stalls has not been required, and the simulators were not built to support it. SME test pilots generally say that legacy simulators, if pushed beyond valid operational range, are much more gentle and forgiving in a deep stall than an actual aircraft.
The new rule requires that simulators accurately reflect an aircraft’s handling characteristics up to 10 deg. beyond the stall AOA for all aircraft, including fly-by-wire systems. Simulation maker and training provide CAE had asked the FAA to reconsider the requirement in some cases since pilots, after recognizing the situation and implementing a recovery procedure, will reduce the AOA before the stall occurs. The FAA disagreed, noting that the 10-deg. threshold will cover most of the pitch excursions the NTSB has seen in accident investigations and that the number is already a recommended practice by International Air Transport Association for simulator aerodynamic modeling.
Expanding the simulator range to 10 deg. past the stall AOA is key to the ultimate cost of implementing the rule because that data may not exist, or may be pricier than the FAA assumed.
For a particular aircraft type, the gold standard for simulator data is provided by airframers and acquired during flight tests, wind tunnel testing or analysis. The FAA concedes, however, that airframers may charge exorbitant amounts of money to collect the information, even if they already have it. In response, the agency says it will allow for third-party models based on a combination of engineering analysis, SME pilot assessment and “improved pre-stall objective testing” using the flight-test data that is already required to be included in the normal flight envelopes of full-motion simulators.
Those extended models may also be “type-representative,” allowing for generic models covering a wider range of aircraft, such as low-wing transports with two rear-mounted engines and a T-tail. An internal FAA simulator study determined such analytically derived, type-representative stall models developed by third-party sources and evaluated by SME pilots could do the job.”
Source; Aviation Week.
