It will be summer before an FAA preliminary rule to upgrade full-motion simulators with extended models to handle full stall training hits the streets, and nearly five years before airlines have to officially put the updated machines to work. Despite the long grace period, the industry is taking a proactive stance on the safety improvements triggered by new flight-training rules, finalized in November and designed to help pilots recognize and recover from fully developed stalls and attitude upsets, often caused by stalls.

“We're finding that there are a variety of different airlines and training organizations trying to do some of this work in lieu of formal requirements,” says Jack Ralston, president of Bihrle Applied Research, a developer of extended simulator models for civil and military aircraft. “A lot of actual users are anxious to get started.” The FAA will issue a preliminary version of the formal requirements, known as the Part 60 rule, this summer, but a final rule could take a year or more to complete.

In the meantime, industry has some fairly solid clues from the FAA in terms of what will be required of simulators in 2019 when the flight-training rule goes live. In January, the agency's National Simulator Program (NSP) published preliminary guidelines to qualify a simulator for the full-stall maneuvers. Simulators today must correctly represent an aircraft in the “approach to stall” regime, where recovery is initiated at the first indication of the stall, typically at the activation of the stall warning system or stickshaker, occurring at or below the stall angle of attack.

The NSP says the 2019 full-stall simulator models should mimic the actual aircraft to “at least” 10 deg. beyond the stall angle of attack and exhibit the same degradation in pitch, roll and yaw stability as the actual aircraft in a stall. Judging how well the simulator performs will be a subject matter expert (SME) pilot “with relevant experience in the aircraft,” typically a test pilot who has performed many stalls.

Guidance on the types of stall models that might qualify comes from an FAA study last year. The agency brought nine SMEs—test pilots with stall experience—to its Boeing 737-800 simulator in Oklahoma City to evaluate stall models of varying complexity, hence cost to develop. The group evaluated the standard as-delivered 737-800 simulator model, but with a more pronounced “buffet,” the rumbling that occurs as airflow separates on the top of the wing near stall; a Bihrle-developed “representative” model built from computational aerodynamics, wind-tunnel data and input from a test pilot with “extensive” stall experience in the 737; and a “specific” model Boeing built from 737-800 flight-test data from “hundreds” of fully developed stalls, including turning stalls. Given the flight-testing required, a specific model will be more costly to develop than the representative model, which could be built to emulate a variety of aircraft with similar characteristics (low- and under-wing engines, and conventional tail). In some cases, flight data may not be possible to obtain for aircraft that are out of production.

On average, the test pilots “somewhat agreed” that any of the models could be used for stall-recovery training, but “wide differences of opinion arose between the particular models,” say the authors of the study, including Jeff Schroeder, FAA chief scientific and technical adviser for flight simulation systems. It could be that more than one model is needed, Schroeder concludes.“Some preferred to be challenged by what could happen in a stall. Others preferred the more benign, typical responses,” he says. “This is an important training question; do you show what could happen or what happens on average? Perhaps the best answer is to show one of each.” The tests confirmed what Birhle's and other experts had suggested, that representative stall models are feasible when flight data are not available.

What the test pilots did largely agree on was that the correct buffet response is important to training. “Even after more than doubling the buffet response [of the unmodified simulator], most of the test pilots believed the buffet cue was less in the simulation than in the aircraft,” says Schroeder.

Along with developing representative models, Birhle also has come up with a technology to upgrade simulators without changing the basic system, an external “stall box” the FAA used for the model testing in Oklahoma City. Ralston says the stall box overwrites the existing simulator model when the aircraft reaches certain thresholds near and beyond the stall, and would not require the basic simulator to repeat its qualification. The system can also drive tablet-based displays of yoke, throttle and attitude positions, and g-force envelopes to help instructors keep pilots within valid bounds of the extended simulator model and to replay pilot performance afterward.

For the stall box, Ralston envisions a two-tiered system where airlines would pay one price without the box active and a higher price for an activated stall box. He says, “You would not have to dedicate a single simulator for full stalls.”