A recent study by the US FAA's Office of Aviation Medicine reveals a "large disparity" between how pilots with aerobatic backgrounds perform compared to those with only ground-based simulator experience when recovering an aircraft from an upset state.
Upsets are linked to loss-of-control events, which the FAA says was the leading cause of hull losses and passenger fatalities in air transport operations worldwide between 1998 and 2007, causing 25% of all crashes and nearly 40% of all fatalities.
"Air transport training programs typically contain a module instructing pilots how to recover an airplane from an upset," the authors of the study say. "However, the effectiveness of such simulator-based training remains uncertain."
The finding is significant given the decreasing numbers of military-trained pilots, who are versed in aerobatic manoeuvres during training, say the authors of the study, which evaluated transfer of upset-recovery knowledge.
"For military trained pilots there are no unusual attitudes, only unexpected attitudes," the report, published in September, concludes. "By contrast, most air transport pilots flying today have never experienced the extreme pitch and bank angles and high G forces associated with severe airplane upsets. Indeed, most have never been upside-down in an airplane even once."
The original premise of the study was that pilots trained in ground-based simulators that can induce continuous G-loads (centrifuge simulators) would fare better during an in-flight test in a single-engine aerobatic aircraft than pilots who had been trained to recover from the same upsets using only a fixed-based desk-top computer. Performance was based on altitude loss, G-forces imparted on the aircraft and other factors.
Airline pilots are traditionally taught upset recoveries on the ground using motion-based hexapod simulators, which can produce certain G-forces for a short duration while moving in three rotational and three translational axes.
For the study, the performance of both groups was compared to a third "control group" that received no training prior to the aerobatic flight. "Although we conducted flight testing in a general aviation airplane, our research has important implications for heavy aircraft upset-recovery trainers," the authors say.
While both the desk-top computer-trained pilots and the G-based centrifuge-trained pilots outperformed the control group, and the G-based group did somewhat better than the desk-top group, what surprised researchers the most was that a separate group of pilots trained in "all-attitude manoeuvring in actual airplane" scored "far smaller" altitude losses when recovering the aircraft in flight trials.
Further, the authors say informal conversations with current airline pilots suggest that "while virtually all regard the company-provided upset training they receive as useful, a significant number also perceive it as a pro forma approach to a serious safety problem - better than nothing but far from what would be desirable if training costs were not a paramount consideration."