UPRT and ICATEE Part Two
Almost ALL of the folks I have had the pleasure of working with in airline training management are fine people with great integrity. For the most part, they provide excellent training, and there is no question on their "desire" to provide the best training. Other examples of negative training in UPRT are due to a lack of knowledge on the training capabilities of the training infrastructure.
The American Airlines Advanced Manuevers Program is a great example of this. It was never their intent than to provide other than world-class quality training. Indeed, we will never know if one of the benefits of this training was in one or several aircraft accident "saves". So I am not here to impugn the developers of this training program. However, we have learned from AA 587, and how limitations in academic understanding, and the practical demonstration of maneuvers in a simulator can have negative consequences.
In-appropriate use of rudder was a finding in this accident. A simulators ability to accurately portray "yaw or side-load heave" as produced by aggressive rudder input is very limited. If the simulator WAS accurate in this motion cueing capability, the sim response to that aggressive input would have been so violent and direct (like the actual aircraft) that you would never have to explain to the pilots why that input was in-appropriate. However, if we combine the lack of realistic g-cueing capabilities in a FFS (particularly those departing the normal flight envelope), and compound this problem with a lack of realistic feedback tools to the pilots and instructors; then we exacerbate the problem. In current FFS, the lack of useful feedback tools can actually REWARD in-appropriate flight control inputs. Using rudder as a roll strategy is VERY effective, however it can produce very negative consequences such as aircraft structural limits being exceeded. ICATEE has identified the necessity of providing enhanced feedback capability to instructors and pilots as a primary need (flight control inputs, flight displays, VN diagram, alpha/beta plots, etc.). The good news is that this is not a formidable technical or cost issue.
This brings up the next challenge; better instructor UPRT training and standardization. In ten years of working as a Check-Airman, I was never given any explicit training in how to teach and evaluate Upsets and Stalls, and certainly in how to use simulators appropriately (and I would consider the bulk of my instructor training and standardization efforts provided by my company as excellent! This seems to be an issue predominately with Upset/Stall training).
Unfortunately, in reviewing our surveys to airlines and operators, this appears to be the norm in the global industry. Many of you may have performed "free-play" with the simulator, expecting that it's accuracy was excellent in all attitudes and maneuvers. Because of this wrongly held notion, too many instructors will provide their own "upset techniques" and a lack of airline training management appreciation on FFS limitations, and the need to ensure UPRT standardization becomes problematic.
One other quick example. Several posters here have identified pre-set maneuvers in simulators in providing upset scenarios. At my airline, we selected one of these maneuvers for our recurrent training cycle in the B-737NG. This was to be a high altitude, heavy weight upset, designed as a TCAS wake encounter that pitched the aircraft up to 20 degrees, and then began an aggressive roll through 130 degrees. So far so good. This was selected from a preset feature on the simulator instructor panel, and the maneuver began!
As the pilot flying, I identified the pitch and roll, cross-checked flight deck instruments, and then initiated a recovery by turning off the A/P and A/T. I provided measure nose down elevator, and as roll began, provided measured aileron input (measured in that flight control inputs were not abrupt, but initiated in respect to the lack of aero-damping at altitude). Imaging my surprise as I reached FULL aileron input that the roll had not reversed, nor even slowed the rolling rate! The jet continued to roll just as quickly past 130 degrees with full aileron input! The next thing I remembered was the instructor yelling at me to feed in rudder!
ICATEE work has made me aware that more damage is incurred during high altitude upsets by pilots inputting improper or overly aggressive flight control inputs. So this selected maneuver needed to be investigated. What we found is a simple and prevalent condition in most level D simulators in use. It is what we can call the "wash-out" effect. When an upset maneuver was pre-selected, the simulator went into an "aerodynamic freeze" whereby no flight control inputs provided by pilots would be honored or recognized by the simulator, until the preselect parameters were achieved (could be 2-6 seconds or so). The net effect of this training exercise was to "teach" our airline pilots that in the event of an high altitude upset, that they would need to take aggressive flight controls inputs. Worse training than if we had done nothing at all!
End of Part Two