Upset Training in the Age of Fly by Wire

 - May 6, 2016, 11:45 AM

The dramatic recovery by the pilots flying a Falcon 7X in May 2011 over Malaysia following a pitch-trim failure and runaway nose-up pitch trim demonstrates an indisputable fact: upset recovery techniques apply to all aircraft types, including those with fly-by-wire flight controls such as the 7X. Many other aircraft types with fly-by-wire controls have entered service or will soon, and this naturally is a subject that pilots who fly these aircraft will want to understand.

According to the Skybrary online safety repository, which is supported by ICAO, the Flight Safety Foundation, UK Flight Safety Committee and European Strategic Safety Initiative, “Some modern types have fly-by-wire primary control systems with built-in protections which prevent the exceedance of critical flight parameters and which can ensure recovery without manual response from the flight crew. However, flight crews must clearly understand that certain system failures can degrade these protections such that the aircraft is left with reduced or no protections and become, in effect, a ‘basic’ aircraft.”

The key point in the above paragraph, that a fly-by-wire airplane can revert to “basic” flight control with little or no envelope protection, is important to understand. While fly-by-wire flight control systems incorporate built-in protections that can help prevent unusual attitudes and loss of control, these systems can also revert to alternate, normal or direct modes in which some or all of the fly-by-wire envelope protections are eliminated, leaving control of the aircraft in a mode that mimics a conventional cable- and/or pushrod-controlled aircraft.

The issue of upset prevention and recovery training (UPRT) for fly-by-wire aircraft has garnered interest, not just because of the 7X incident but also as a result of airline accidents. One was the June 1, 2009 Air France Flight 447 accident, in which the pilots stalled an A330 into the Atlantic Ocean, killing 228 people. The other was the Dec. 28, 2014 crash of an A320, Air Asia Flight 8501, in which 155 people were killed, following a stall and loss of control.

According to the Air Asia Flight 8501 accident report by Malaysia’s Komite Nasional Keselamatan Transportasi, “The upset recovery training was included in the aircraft operator’s training manual. The aircraft operator advised the KNKT that the flight crew had not [undergone] the upset recovery training on the Airbus A320, and this referred to FCTM [flight crew training manual] Operational Philosophy: ‘The effectiveness of fly-by-wire architecture, and the existence of control laws, eliminates the need for upset recovery maneuvers to be trained on protected Airbus aircraft.’”

However, Airbus does provide recommended UPRT training techniques in simulators for pilots undergoing initial type rating and recurrent training in all Airbus models (A318 through A380). In an Operations Training Transmission (OTT) issued on March 10 last year, Airbus wrote: “This document is intended to guide operators and Approved Training Organisations (ATOs) in the conduct of Upset Prevention and Recovery Training (UPRT). Further documents are under development by Airbus and will be released in due course. The scope of this document is to address UPRT conducted in a Flight Simulation Training Device (FSTD) qualified for the purpose, during type rating and recurrent training.”

David Owens, Airbus senior director of training policy, told AIN that “the intent behind the statement is still valid,” regarding the FCTM quote in the KNKT accident report. But he added that it is an old statement and that Airbus’s current philosophy is reflected in the OTT mentioned above and other Airbus documents. Essentially, Airbus doesn’t recommend dynamic upset training in simulators, he explained, “because the motion cues are not accurate.” Airbus doesn’t want pilots to learn something in the simulator that doesn’t safely mirror the actual airplane’s characteristics. What Airbus would prefer is that pilots learn the dynamics of upsets in their initial licensing in an actual airplane, which is reflected in upcoming EASA regulations (see below).

“Upset training should be done before a pilot gets near a commercial aircraft,” he said. This philosophy has to do with the way new pilots are currently trained. “When I went through basic training 25 years ago,” he said, “I did hour after hour of unusual attitudes, IFR and VFR. That’s no longer part of basic training for airline pilots. When we had mostly military pilots [hiring on as new airline pilots], the majority had been through that. It’s very different now when pilots go on the line flying the A320, but now we’re going to change how we do this.”

With regard to simulators, Airbus is all for using simulators for a specific portion of upset training and that is the recognition phase. “In [those stall accident] cases, we had pilots that didn’t recognize the situation,” Owens said. “What we want to push is the concept of accurate recognition. Think about each event [the Air France and Air Asia accidents]: the pilots didn’t know what was happening. Recognition is the root cause, so we’re focusing on recognition.”

Randall Brooks, vice president of training and business development for upset-training provider Aviation Performance Solutions (APS), explained that it shouldn’t matter what kind of flight controls are involved. “Most UPRT concepts are 80 to 90 percent across all types, no matter whether it’s a propeller or jet airplane,” he said. APS teaches UPRT in the Extra 300 (a two-seat aerobatic piston single) and also in regional jet full-flight simulators.

He pointed to the Falcon 7X runaway trim recovery as a “perfect example of why [UPRT] can be beneficial to a pilot.” The 7X crew used a banking maneuver to lower the nose from the resulting nose-high, low-speed upset. “He was a former fighter pilot,” Brooks said, “and he knew how to maneuver [the airplane] to recover it. It’s not required in civilian licensing, but it’s the kind of thing we’re teaching in our course.” (Of course other upset-training providers teach that maneuver as well.)

“That’s 75 percent of the argument,” he added, “but the other part is that there are real manual handling skills and benefits [to UPRT]. Our typical on-aircraft training is four flights, from three-and-a-half to four hours. Pilots who fly with us or other [training organizations] are going to learn more in that period than any similar three to four hours of instruction they ever have in their careers. It’s expensive and not required by regulations, but it is highly valuable in terms of reducing pilots’ exposure to the number-one killer in aviation. No other four hours you could spend statistically have that same benefit.”

APS frequently works with Airbus pilots and also trains instructors who then teach their own pilots who fly Airbuses. “We have to have a working knowledge of Airbus control laws and so on,” Brooks said. “In any of these upsets, we assume that the flight control laws are degraded. Our point is that when the flight control laws are degraded, it’s just like any other airplane. We had a captain from an international airline tell me that an Airbus can’t stall, and I’ve had to remind him where it has. Things don’t always work out. We like redundancy, in engines, [electronics] busses and so on. I don’t understand why it is we don’t appreciate the redundancy of UPRT. If things get out of the normal envelope where we normally fly, having those skills provides an additional safety net when the flight control laws don’t work, when the controls jam. We have more confidence, and we have the manual handling skills.

“The other problem,” he added, “is the assumption of a lot of people who are trained only to 60 degrees of bank and 30 degrees of pitch; they aren’t aware that when they get beyond those boundaries, separate skills are required, and many of those skills are counterintuitive. They aren’t hard, but beyond the normal envelope, some of the normal skills don’t apply. When you pull back on the controls, the nose goes up and the airplane slows down; that’s true 100 percent in the normal flight envelope. But in a 95-degree bank, all of a sudden everything that has worked all my life no longer applies. To think you can’t get there in a fly-by-wire-protected aircraft defies reality. Air Asia 8501 got to a 100-degree bank at some point, and it got somebody facing a situation that traditional licensing [didn’t cover].”

New Training Requirements

There is a regulatory push to provide more UPRT to pilots, including upcoming new EASA rules (based on ICAO recommendations) that will require UPRT in aircraft type rating and initial training as well as new FAA rules for simulators that will require expansion of aerodynamic modeling. The new FAA rules will require support for the following training tasks: stall and stick pusher, upset recognition and recovery, engine and airframe icing, takeoff and landing in gusting crosswinds and bounced landing training.

The new EASA training rules also require three hours of in-aircraft aerobatic training starting in 2018, with an appropriately rated instructor, although this applies only to pilots with commercial or ATP certificates flying multi-crew airplanes, according to Brooks.

“One thing is clear,” he said. “Instructors delivering this training should have received some training in UPRT themselves. How are they supposed to teach it if they’ve never had any experience? It’s the blind leading the blind. Now, people don’t even know what UPRT is. If they go to [training at a major training provider], they have a couple of tasks–unusual attitudes and stalls. In real life, unusual attitudes and stalls can happen at the same time. Training in unusual attitudes and stalls separately doesn’t train you for when it happens simultaneously.”

Brooks cited statistics that show that loss of control ranks twice as high as any other accident causal factor. “There’s a reason for that,” he said. In an analysis of 20 loss-of-control accidents between 2001 and 2010 done by Brooks and APS president BJ Ransbury, they found four assumptions that proved to be incorrect (in 16 of the accidents that provided valid data for the analysis).

The first assumption was that the airplane will mostly remain inside the normal envelope, but this wasn’t the case in 60 percent of those 16 accidents.

The second was that existing training would be adequate to prevent these accidents, but in more than 60 percent of the accidents, he noted, “it required some skills that the pilots were not provided or didn’t show proficiency with during normal licensing training.”

The third assumption was that pilots would be able to make sense of all the warnings, cautions and advisories that might occur during an upset. “In cockpit voice transcripts,” he explained, “[pilots] tell you: ‘I don’t understand what’s going on.’”

Finally, he said, “In 100 percent of the cases the pilots did something not consistent with what their training should have told them to do.” A key example was not lowering the angle-of-attack during a stall.

All this highlights, he said, “that existing training leaves a deficiency for pilots. It does not prepare pilots for what they need to do in a loss-of-control event. This situation is going to remain until we grasp that and change training. That’s what the ICAO recommendations and EASA rule changes intended: to provide pilots with these skills. It won’t eliminate loss-of-control accidents, but if we cut them by half, we will make a dent.”

None of this is news to aviation safety experts. Don Bateman, the former Honeywell chief engineer flight safety technology (now retired) who was the inventor of the life-saving ground proximity warning system, wrote a paper that was presented to the American Institute of Aeronautics and Astronautics Guidance, Navigation and Control Conference in 2010.

The paper, titled “Some Thoughts on Reducing the Risk of Aircraft Loss of Control,” addresses the difference in risk for loss-of-control (LOC) accidents between conventional and fly-by-wire (FBW) aircraft. “The LOC risk is highest for the conventional ‘pulley and cable’ control system aircraft. While FBW-designed aircraft have demonstrated significantly lower risk when compared to pure conventional ‘pulley and cable’ aircraft, FBW aircraft are not immune to LOC.” That said, the historical loss record is approximately one pulley-and-cable aircraft for every 3.7 million departures versus one FBW aircraft loss per 37 million departures, so clearly there is a benefit to FBW architecture.

Manufacturer Recommendations

AIN asked business jet manufacturers for their recommendations regarding UPRT.

Gulfstream pointed to its partnership with FlightSafety International in helping the simulator manufacturer reprogram some of its Gulfstream simulators to better model the aerodynamics outside the normal flight envelope. FlightSafety is using these simulators for a one-day UPRT course where students are able to practice recovering from real accident scenarios and experience the simulated results when the recovery measures fail. For its fly-by-wire G650, Gulfstream is recommending that pilots take the one-day FlightSafety course, which should be available for that airplane at some point.

Dassault Falcon explained that it publishes upset and stall recovery procedures. “These and a multitude of possible abnormal scenarios are covered within our Abnormal Check Lists, Crew Operational Documentation Dassault EASy and AFM. All those procedures are tailored by our flight-test pilots.”

The manufacturer also pointed out that at the end of last year it issued guidance to Dassault-approved training providers on improvements to UPRT training for pilots flying under EASA and FAA type ratings. “This guidance is not about what to train but how to improve the existing training material.

“Our goals: Ensure the trainees have a comprehensive knowledge of unusual positions as a phenomenon and associated risks and be able to identify any diverging situation; train the trainers; optimize the ground and simulator courses; adhere to and accomplish the OEM recovery procedures.”

The banking recovery maneuver was added to Dassault’s upset recovery training program in late 2014, according to Frédéric Leboeuf, v-p of Falcon operational support directorate.

The SuperJet International Training Department provided a detailed answer to AIN’s questions, applicable to the fly-by-wire RRJ 95 (Superjet 100), summarized below:

The RRJ 95 (Superjet 100) is fly-by-wire, with full flight envelope protection [and] sidestick control. The flight control laws are two: normal or direct (in case of multiple malfunctions).

An angle-of-attack indicator is always displayed on both primary flight displays.

These factors were considered when, at [the] very beginning of [the] type rating course (in 2011), maneuvers and exercises concerning upset recovery, stall recovery and unreliable airspeed indications were introduced in the training syllabi.

All the technical literature developed from worldwide organizations was considered (i.e. stall recovery technique).

Training on [the] FFS [full-flight simulator] was focused mainly on:

• Demonstration of flight-envelope protection (in normal mode) in different configurations and altitudes, its limitations, risk of complacency.

• Series of stall events (in direct mode) in different configurations and altitude.

• Series of upset recoveries in different configurations and altitude.

• Different aircraft behavior in normal and direct mode (i.e. auto-trim capability).

• Use of AoA indicator (most pilots are not familiar with it because it is not common on airliners).

• No feedback from sidestick, either for input coming from [the] autopilot or from [the] other pilot.

•Emphasis on crew coordination/integration during stall event or upset recovery; function of priority pushbutton and results of dual stick input on sidestick.

Besides these and other considerations, [the] limits of simulations were emphasized because simulators are certified to “approach to stall” [because] the flight envelope has been demonstrated [only] till this phase.

Embraer did not provide a response to AIN.

A Boeing spokesman declined to answer AIN’s questions and said, “I’ll have to point you to former pilots or third-party experts such as the Flight Safety Foundation.”

The Flight Safety Foundation has addressed this subject, most tellingly in an article about the safety organization’s 23rd annual European Aviation Safety Seminar held in March 2011 in Istanbul. J.A. Donoghue, writing about stall prevention and recovery training in AeroSafety World magazine, quoted Claude Lelaie, special adviser to the Airbus president and COO, who “cut to the heart of the remedy for pilots finding themselves in a stall or near-stall condition: ‘If you push on the stick, you will fly!’”


apstrainingusa's picture

Despite steady improvement in the reliability of engines, pilots still routinely practice procedures required to correctly respond to engine failures. This training creates redundancy through training, which allows pilots to handle the rare powerplant problem even when encountered at the most critical times, such as just prior to takeoff rotation.

Although the failure of digital flight control systems (“fly-by-wire”) are highly unlikely, history has taught as that simple issues like pitot tube icing and other events can compromise their protective features, and in some cases render them ineffective. Applying the same principle of redundancy through training (providing pilots with necessary back up skills) as we use in the case of engine failures, we should teach pilots how to handle their aircraft in the case of flight control system failures; regardless of how sophisticated the control system. Why don’t we?

All civil licensing training is based on the false premise that aircraft will always remain inside of their ‘normal’ flight envelope. This is evidenced by the fact that no current licensing training for Commercial pilots is required beyond 60 degrees of bank or more than 30 degrees nose high or low. While such aircraft upset situations beyond these thresholds are rare, they are often catastrophic. Only 2% of accidents are due to loss of control in-flight (LOC-I) but they result in approximately 40% of all airline fatalities.

LOC-I has become the leading cause of fatalities in every category of aviation. This can be significantly reduced through focussed Upset Prevention and Recovery Training (UPRT). Until we recognize that current training does not adequately prepare pilots to appropriately respond when outside of normal operating parameters (which is highly counterintuitive) and require comprehensive UPRT, the loss of lives due to LOC-I will continue. Moreover, the psycho-physiological aspects of airplane upsets that impact the performance effectiveness of crews can not be addressed in today's flight simulators, even those with extended envelopes. For this reason, as Capt. Owens points out at a high level, the November 2014 ICAO Manual on Aeroplane UPRT (ICAO Doc 10011) cites industry-compliant on-aircraft UPRT as being essential to all pilots to maximize their ability to access skills and knowledge in a time-critical, life-threatening crisis such as an airplane upset event.