by Palmer, Bill
It is possible they were too tired to be performing well.
The captain had allegedly stated at 01:04, "I didn’t sleep enough last night. One hour – it’s not enough."
About an hour before the loss of airspeed event, the captain offered First Officer Bonin an opportunity to take a nap. One must assume that he had some reason to offer this, such as him looking or behaving tired. The captain said, “Try maybe to sleep twenty minutes when he comes back or before if you want.” Bonin turned down the offer, “Oh … that’s kind” he said, “For the moment I don’t feel like it, but if I do feel like it, yeah.”
One might tend to conclude that Bonin clearly said he was not tired. After all they had only pushed back three hours earlier. But the captain followed up with “It’ll be a lot for you.” Thus apparently knowing and trying to convince Bonin that he had already had a long day and he was probably not well rested.
When Robert returned from his break and described his ability to sleep on his break as “so-so,” he asked Bonin if he was OK. He probably had a reason to ask.
It has been postulated that once the A/P disconnected the issue of fatigue would have been overridden by the adrenaline of the moment. Maybe it did to a degree, to make them awake, and to give more strength to pull back on the sidestick, perhaps without even knowing it. This is the source of strength that allows mothers to inexplicably lift cars off of their children. But I do not think it is the source for better mental performance.
It is during these high stress events that we read of soldiers, fire fighters, police, and other situations of high stress where they perform under this stress not by working out the proper solution with clear thought, but that their training ‘kicked in.’
As mammals, our hard-wired reactions to stressful situations are to freeze, run away, or fight back. Thinking of a brilliant solution to the problem at hand is usually not in that mix. The pilot flying needed to know to push forward on the stick, not more strength to pull back.
The two first officers may have been affected by a degree of combat stress. Combat stress has been defined as “The perception of an imminent threat of serious personal injury or death, or the stress of being tasked with the responsibility to protect another party from imminent serious injury or death, under conditions where response time is minimal.35” I think it is clear that much or all of those conditions applied in this case.
Tunnel vision, auditory exclusion, the loss of fine and complex motor control, irrational behavior, and the inability to think clearly have all been observed as byproducts of combat stress. Consider the multiple alarms sounding (autopilot disconnect cavalry charge, stall warning, chimes, and the constant sounding of the C-chord). It is possible or likely that these were tuned out (i.e., auditory exclusion)? Bonin’s overcontrolling the roll and pitch inputs despite admonitions from Robert to the contrary, build a case for the loss of fine motor control.
In 1950, S.L.A. Marshall's The Soldier's Load and the Mobility of a Nation was one of the first studies to identify how combat performance deteriorates when soldiers are exposed to combat stress. Marshall concluded that we must reject the superstition that under danger men can be expected to have more than their normal powers, and that they will outdo their best efforts simply because their lives are in danger. In many ways the reality was found to be just the opposite, and individuals under stress are far less capable of doing anything other than blindly running from or charging toward a threat. Humans have three primary survival systems: vision, cognitive processing, and motor skill performance. Under stress, all three break down.
When Bonin’s training ‘kicked in,’ it seems likely that the only training he had to deal with the stall warning was the application of TOGA power. When that failed, they were out of ideas and there was no mental capacity to reason out a new one. The next level down of instinctual reactions from him may have been to simply pull back on the stick.
Studies show that 17 hours awake is equivalent to .05 alcohol level. The combination of the startle effect with a diminished mental capacity due to fatigue and stress, during a time when there was a need to think quickly and accurately, and/or call upon past training that may not have been adequate for this situation, may have led to the failure to maintain level flight initially, and then to the apparent confusion which inhibited their ability to comprehend and recover the airplane.
Training
Many comments have been made on various forums concerning the training aspect.
One of the factors missing in training and pointed out in the final report’s recommendations36, related to the lack of “specific and regular exercises dedicated to manual aircraft handling of approach to stall and stall recovery at high altitude.”
All three pilots had their A330 and A340 training in the context of an additional rating, taking into account their previous A320 experience, and based only on the differences between the types. As a result none had done any stall training in a A330 or A340 simulator, because the protections, indications, and warnings between these three aircraft are virtually identical.
Stall training, not only at Air France but industry wide prior to the accident, concentrated on stall recognition and recovery at low altitude, where those incidents were considered most likely to occur. Even though there had been stall accidents originating at high altitude (as previously noted).
While the principles of stall recovery are similar in both altitude regimes. Significant differences between low and high altitude do exist:
Significantly less excess power is available at high altitude, requiring recovery primarily with pitch for angle of attack reduction.
A lower stall angle of attack at the Mach numbers experienced with high altitude flight, and consequently a narrower stall margin than at low altitude.
Stall training, however, focuses on prevention and recovery before the actual stall is encountered, or “incipient” stall recovery.
In stall training exercises, students recover at the first indication of a stall, that is the stall warning or perceived buffet. The simulator training is not carried into the full stall scenario, nor is the simulator designed to do so.
Prior to the accident most stall recovery training also focused on minimal loss of altitude in the recovery. Rapid application of full power was the initial action, and at low altitude the application of full power often solved the angle of attack problem by itself. Very little pitch down was required to make the recovery and altitude loss was minimized. But again this was recovery from an incipient stall, not a full stall. Pilots were taught to take action as soon as the stall warning or other signs of stall presented themselves. Also keep in mind that stall training in the simulator begins with “OK, we are going to do some stall training now,” and it is not presented in the environment encountered by this crew.
Pilots are also taught that in Alternate and Direct laws the flight envelope protections are lost (as stated on the ECAM when the control law degrades), and that one must be more careful in handling the airplane. However, in the case of AF447, the stall warning first sounded momentarily at 02:10:10, almost immediately after First Officer Bonin inexplicably more than tripled the pitch attitude to 11°. It was silent for 38 seconds then sounded at least four times before TOGA was selected at 02:10:55. While some nose-down inputs were made, the pitch attitude continued to increase and sufficient stick input to establish a nose-down recovery attitude was not made.
The ECAM presented: “Alternate Law, protections lost.” First Officer Robert only got as far in the ECAM procedure to say “Alternate Law protections, lo…”. The transcript reads “Alternate Law protections (law/low/lo),” indicating that the transcriber did not understand what Robert was trying to say. This message is one that would need to be relayed clearly to the pilot flying, and there would need to be assurance that he understood it, as it affects the way he flies the airplane. At the same time that Robert was reading that step, Bonin was asking him about the previous one, “engine lever?” That is becaus
e Robert’s reading of that step was also unclear. So, it is not clear that Bonin, the pilot flying ever really understood that the airplane was in Alternate Law and that the flight envelope protections were lost.
During certification flight tests, the real airplane was not fully stalled at high altitude, and therefore data on aircraft behavior and performance beyond that was non-existent. It is impossible to create a realistic simulator model from data that does not exist. I have performed a high altitude full stall scenario in A330 simulators by two different manufacturers and each one behaved somewhat differently. In those simulators it took about 10,000 feet to recover from the fully stalled situation at high altitude. I have no idea of the accuracy of this in relation to the actual airplane’s ability to recover from an extreme fully stalled situation at high altitude.
In later tests by Airbus, angles of attack up to 14° were reached with significant buffeting, and that was proved to be recoverable. The point at which it might not be recoverable is an unknown. It would come as no surprise to many if the airplane was not recoverable at all from angles of attack in excess of 40°. But during the flight, even though extremely high angles of attack were reached (up to 60°) whenever the sidestick was positioned forward, the nose pitched down and the angle of attack reduced. How much altitude it would take to complete a recovery is anyone’s guess, but it would be many thousands of feet.
There is no training in spin recovery of airliners either. A spin is a result of an asymmetrical stall. The entire focus of training is to prevent those scenarios. In my opinion, AF447 may very likely have ended up in a spin had it not been for an active automatic yaw damper working furiously throughout the remainder of the flight to counteract yawing motion. A spin would have resulted in an even greater descent rate and a much lower likelihood of recovery on instruments alone.
There is an old saying that says, “A good pilot uses his superior knowledge to avoid situations that require his superior skill.” Acquiring that superior knowledge is the trick. Memorizing the facts and answering the test questions does not always develop insight. In fact, it rarely does.
A pilot needs to understand the airplane, in all its modes. But that level of knowledge, understanding, and insight can be difficult to acquire in a short amount of time. There is a huge volume of material to be assimilated.
Like a surgeon that learns a new procedure, one does not become an instant expert. Once simulator training is complete and the new pilot is certified to fly the airplane, his or her initial experience in the airplane (15-25 hours typically) is with a line instructor, on actual passenger carrying flights. Most of that time is spent in cruise, and on a long-range airplane such as the A330, there will most likely only be a few takeoffs and landings with the instructor present before he is satisfied that the new pilot is “good to go.” But the learning process is far from over.
The new pilot will be restricted slightly in the weather that an approach may be made in, and he will not be authorized to fly with another pilot who also has fewer than 75 hours in the airplane. It will be up to the pilot’s own initiative as to how well he reviews his previous materials, and refreshes his memory of all the material that he had to try and absorb quickly just a few weeks prior.
It can seem overwhelming at first, so much to remember, so much to look for, so much to keep in mind. Until it clicks, and then the airplane is a beautiful friend to fly. But it will take a conscious effort to do it. If one only thinks about the routine functions used on every flight, some of the lesser used functions and nuances can be forgotten.
Everything works so well, so much of the time, that when things are not going right, it is easy to fall into the trap of assuming that the automation is doing the right thing. For example, that anytime the flight directors are on, that they are worth following.
There are red disconnect buttons on the sidestick (for the autopilot) and on the thrust levers (for the autothrust) that give the pilot instant control over the airplane. I have told my students many times, “If it’s not doing what you want it to do, make it do it! Click it off, push it over, make it turn, whatever it is you want to happen, Make it Happen!” Of course, you must be thinking far enough ahead of the airplane to know what that is.
The Airline Transport Pilot (ATP) standards for the US dictate that the successful candidate must be the “master of the airplane.” It is not just a skill set, it is also a mind set.
The Final accident report by the BEA states the following as a component of the cause of the accident:
The occurrence of the failure in the context of flight in cruise completely surprised the pilots of flight AF447. The apparent difficulties with airplane handling at high altitude in turbulence led to excessive handling inputs in roll and a sharp nose-up input by the PF. The destabilization that resulted from the climbing flight path and the evolution in the pitch attitude and vertical speed was added to the erroneous airspeed indications and ECAM messages, which did not help with the diagnosis.
The crew, progressively becoming de-structured, likely never understood that it was faced with a “simple” loss of three sources of airspeed information. In the minute that followed the autopilot disconnection, the failure of the attempts to understand the situation and the de-structuring of crew cooperation fed on each other until the total loss of cognitive control of the situation.
The airplane went into a sustained stall, signaled by the stall warning and strong buffet. Despite these persistent symptoms, the crew never understood that they were stalling and consequently never applied a recovery maneuver. The combination of the ergonomics of the warning design, the conditions in which airline pilots are trained and exposed to stalls during their professional training and the process of recurrent training does not generate the expected behavior in any acceptable reliable way.
In its current form, recognizing the stall warning, even associated with buffet, supposes that the crew accords a minimum level of “legitimacy” to it. This then supposes sufficient previous experience of stalls, a minimum of cognitive availability and understanding of the situation, knowledge of the airplane (and its protection modes) and its flight physics. An examination of the current training for airline pilots does not, in general, provide convincing indications of the building and maintenance of the associated skills.
In the absence of reliable speed indication, an understanding of the physics of high altitude flying, gained through training in the fundamental principles of energy conversion, equilibriums of forces, and lift and propulsion ceilings, could have considerably helped the pilots to anticipate the rapid deterioration in their situation and to take the appropriate corrective measure in time: initiate a descent.
The final report theorizes that over-speed was a strong risk in the PF’s mind. I do not disagree, as several statements were made about having a “crazy speed,” and at one point he extended the speed brakes. This was the consequence of the fact that, in theoretical teaching (notably Airline Transport Certificate), the risk of “high speed stall” is presented equally with the more classic “low speed stall”. Though low-speed buffet is quite well known to pilots, excursions beyond maximum speed limits are not demonstrated in training. Furthermore, vibrations (linked to buffet) were erroneously associated with over-speed.
Air France’s Aeronautical Manual describes in great detail, over 38 pages, the physics of high-altitude flight with real cases. This knowledge is also included in the theoretical teaching that is supposed to be provided at an advanced stage in the training of a future airline pilot (Airline Transport License theory, type rating performance). The climbing flight path that was initially more or less deliberate on the part of the crew is likely a clue to the insufficient assimilation of these theoretical notions.
Modern aircraft such as the A330 are far less critical in the transonic range than their earlier generation counterparts. In those airplanes, an over-speed condition could lead to an aft shift of the center of lift, interruption of the airflow over the tail, and an unc
ontrollable dive, known as Mach tuck.
Unfortunately, the characteristics of exceeding the maximum speeds for particular aircraft types, and therefore the applicable risk for each, are not well known to pilots.
The Future of Training
Training has certainly improved with time, and I am confident it will continue to do so. There are areas that are subject to improvement, and I have no doubt that progress will be made along these lines.
A current threat is that a large amount of time is dedicated to simply learning to operate the particular aircraft and its technology, along with the company’s procedures for the normal and non-normal operations as required. Little time is dedicated to flying skills, which obviously must be the foundation upon which all the procedural and technology training is based.
Pilots come from a wide variety of backgrounds. It is difficult to quantify what actual experience and skill any one pilot has in handling unexpected circumstances where control of the airplane is at stake. Even if a pilot had the experience and skill when he was hired, how well are those skills preserved when flying on the autopilot 99% of the time for decades?
Loss of Control In-flight (LOC-I) remains the leading cause of accidents over the last 20 years. Other accident causal factors have been reduced due to more reliable equipment and systems providing improved protection and recovery from wind shear, traffic, and terrain threats. Yet the LOC-I accident rate remains virtually unimproved and has been assuming an increasing percentage of the cause of fatalities. According to a Boeing study, there were 1493 fatalities due to LOC-I, one due to engine failure, and 225 due to non-power-plant systems failures (less than 1/6 th of the LOC-I rate).37 Yet considerably more time is spent on those failure scenarios then recovery and prevention of loss-of-control events.