OK, watched the episode, will attempt to answer questions. Usual disclaimer that my expertise, such as it is, is in smaller aircraft and any big iron pilots please do feel free to correct any errors I might make or add your own insights. I am strictly an amateur analyst, and conclusions can and probably will change with more information. What follows is speculation and unofficial at that. I can be right and I can be wrong, so here’s a grain of salt in advance.
I remember at the time that there was much mention of big storms reaching to 50,000 feet. That is an extremely high intensity storm, on par with the one that recently spawned all those tornadoes down south and caused so much damage. I hasten to add that such storms don’t always spawn tornadoes, but they are all very dangerous weather events that any pilot with half a brain wants to avoid. My friends who fly airliners have commented in the past about the limitations of on-board radar in a manner consistent with what is described here.
Keep in mind they were over the middle of the ocean – over land there are typically many more observers and sources of weather information. Due to limited communication there is less opportunity for pilots to either ask about severe weather, report severe weather, or be cautioned by air traffic about severe weather.
In short, they explained how equipment limitations can result in competent, alert pilots suddenly finding all hell breaking loose outside the aircraft. If you don’t know the storm is there you can’t avoid it. You need the information before you can act on it.
I thought they explained the problem of pitot icing quite well.
I do have a quibble with the way they continually described the automatic systems as having “failed”. Yes, sometimes they do just quit working, but they never mentioned that an automatic pilot shutdown is not always a failure. They are designed to shut down under certain conditions. If they do so in response to those conditions is it NOT a failure! In other words, I prefer your phrasing of “disengaged” much better.
You see, for routine flying machines actually do a better job than people. We pilots don’t always like to admit that, but it’s true. For routine flying. If something goes really wrong, though, people are better problem solvers than computers. If flight conditions exceed certain parameters the autopilot (and perhaps other systems) shut down because under those circumstances it’s probably a situation a human will handle better than a machine. You shut down the automatic stuff so it won’t interfere with what the human is trying to do, which might be radically different than what you’d do under normal circumstances. There have been a number of incidents where this worked very well, with either a high number of survivors or even everyone reaching the ground unhurt even with a severely damaged airplane.
I can’t say from the information portrayed if all those systems shutting down was a fault or a planned feature for sure. I can describe some possibilities, but it would take flight information from the data recorders to make an accurate assessment.
Explaining stalls can be difficult. I thought they did a reasonable job, but understanding them can be difficult for the non-flyer, and even a substantial number of pilots have less than ideal understanding of them.
The important point is that flight depends upon generating lift, and lift is generated by a particular sort of airflow over the wings. If that airflow is disrupted you lose lift. Lose enough lift you fall out of the sky. The fact that Flight 447 “pancaked” into the ocean (based upon damage to the recovered debris) leads anyone interested in aviation accidents to wonder about a stall.
Now, about those stalls – pilots get training in stalls early and often. I think I started stall training around hour 3 myself. Every pilot learns about stalls and stall recovery. The basic concepts apply to all airplanes. To pilots, they aren’t exotic. In fact, some people even do them for fun.
:eek:
No, really, let me explain. In training, in good weather, you go up to a sufficiently safe altitude and your instructor deliberately stalls the airplane, then recovers so you can see what it looks like. Then you practice doing it yourself. You practice variations and different scenarios until your reactions become automatic. There’s even something called a “falling leaf” stall where you stall the airplane then hold it in that state for a bit, undergoing an interesting rate of descent where, yes the wings have a bit of a “falling leaf” sort of movement. You practice stalls while in turns, while descending and ascending and while turning and changing altitude at the same time. You do all this so you know how to handle the situation, and what it feels like just before you stall. Mind you, that’s just for private pilot level, and stall practice continues forever in aviation. The problem isn’t stall practice, it’s when an unexpected one leaps up and bites you in the ass when you’re dealing with 1,000 other major concerns.
Now, stalls actually aren’t a matter of airspeed (but there’s a reason for the confusion). It’s a matter of the angle at which the wing meets the air, called “angle of attack”. However, angle of attack is also related to airspeed, hence the reason we talk about things like “stall speeds”. The airspeed can give you information about the angle of attack, which makes it a useful reference for avoiding stalls, but it is not perfect. Stalls can be affected by engine power, the shape of the wing, and so on. This is why, absent airspeed information, setting a specific pitch and power setting also works – because pitch and power affects angle of attack, too.
Another thing to know about modern airplanes at pretty much all levels is that they have been engineered to stall in a docile manner (some fighter and stunt planes aside). The wings are constructed so they won’t lose lift along the entire lengthy all at once – you lose some lift, not all of it. Between the rudder and proper training the pilot need not lose ALL control, you can retain some. And airplanes are designed so that if lift is lost the nose will tend to come down all by itself. That is based on weight and balance and shape, it is not dependent on engine power or flight controls. It’s physics that cause the nose to lower and thus reduces the angle of attack so the plane resumes flying. Not that you want to depend wholly on that to save your life, but it makes the task of stall recovery much easier. The important thing to note is that the airplane “wants” to resume flying, it will assist the pilot in recovery, or at least not work against him.
Stalls aren’t inherently fatal. If you don’t make a recovery they can and will kill you, but recovery is possible most of the time, and pilots are trained in stall recovery from the very beginning. It’s not something you want to have happen to an airliner, it could result in a brown underwear event for those aboard, but it’s not inherently fatal. Time was, before simulators, airline pilots would be required to regularly take up an (empty) airliner and deliberately stall it, and recover, to demonstrate they were still proficient at this. I’m told this is now down in simulators mainly due to cost reasons, as flying empty airliners is expensive. So, bottom line, stalls don’t have to be fatal in a big airplane. They aren’t necessarily violent in a big airplane, but it would have to be a pilot familiar with airliners who describes their stall characteristics in more detail, as, to the best of my knowledge, I have never been in a stalling airliner. Yes, I am hinting it might be possible to be in a stall and not know it - there are some small airplanes I fly with such docile stall characteristics that I can, if circumstances are right, put it into a stall and recovery such that a passenger may be entirely unaware that a stall has occurred (well, except for the audible warning). Could a stall be that gentle in an airliner? I don’t know. On the other hand, given the circumstances, if Flight 447 did stall then crash I’m guessing it wasn’t that gentle. People on board would have known something was wrong, even if they didn’t know what. If a stall turns into a spin it would get REALLY unpleasant.
Take away here - stalls need to be dealt with promptly. They can range from gentle non-events to the horrific, depending on circumstances, but all need to be recovered from promptly.
Anyhow, the consensus view is that Flight 447 was making a routine flight at night from Brazil to France. That is, in fact, routine these days. They had some weather up ahead, and I’m sure they had an eye on it. At a certain point they drop off Brazil radar and are out of radio range with land. That is also routine, particularly in bad weather you can get communication problems, and I’m sure they just continued flying.
Then – whoopsie! - turns out there’s a MONSTER thunderstorm past that little storm that just flew through. Now they have to deal with it. Keep in mind, it’s night, it’s overcast, and they’re over the ocean. The windows will be pitch black, except for an occasional lightning-flash and the ONLY information they get about what’s going are is from the various instruments on the airplane. But the are trained to deal with bad weather and zero visibility. There is nothing to do but keep flying to better skies.
Then the automatic systems start shutting down, alarms go off – but the pilots are trained to keep flying no matter what. I’m sure they were tense, perhaps even frightened, but they’re going to keep trying.
Meanwhile… they have no reliable airspeed information. This is not good. Oh, they’re probably thinking (if they have a second to spare) that the pitot tubes have iced up, it’s not an unknown problem. But there’s no doubt a lot going on in the cockpit.
Alright, let’s talk about cruising speeds in aircraft. Airliners don’t fly at the fastest speed, they aim for an efficient speed, the maximum distance for the minimum fuel, with maybe some concern for speed. That’s one reason for flying at 35,000 feet, the air is thinner and it requires less energy to fly through it. Another consequence is that, yes, at cruising altitude you aren’t very much faster than stall speed. Now, airliner pilots are held to extremely precise flying, and this sort of thing is quite routine. It’s not a safety issue. Landings are closer to stall speeds than most people would imagine, too but really, landing faster isn’t any safer, particularly on runways of limited length. Take-offs are “slow” in a sense, too, but that’s because more of the energy is going into lifting you up, not pushing you forward, and you’re safer getting away from the ground faster. There is a comfortable margin above a stall, even if it sounds small to the non-aviator.
Now, a stall warning is a device that alerts the pilot that hey, you are really getting close to a stall, do something about it (usually – I’ll mention an exception shortly). It might be a buzzer or bell or flashing light or, in the case of airliners, it physically shakes the stick in the pilot’s hand. You’re still not stalled when it goes off, it’s a warning, and there’s still a margin between warning an actual stall. I mentioned an exception, right? In a small airplane, if you’re landing in a rough field you might opt to land at a speed at just the point the warning goes off, because in a rought field you want to land at the slowest speed that is safe, that is, controllable. At the point the warning goes off you are STILL in control of the airplane, but any slower your control degrades.
OK, we know that around 2:10 to 2:14 am the automatic pilot and some other systems shut down and the pitot system is not functioning. The airplane has no reliable airspeed information, without which the machine can’t fly itself, so it’s turning everything over to the bald apes in the cockpit. We know there’s a storm, so they almost certainly were in turbulence and getting bumps and jolts, which isn’t any fun and doesn’t help. It’s probably ice on the tubes, but really, for the guys in the front office why doesn’t matter at this point, they need to fly the airplane, and their information is limited. They can’t look outside to confirm anything, it’s pitch black. So they have to rely on the instruments… which they already know aren’t fully functional.
One of two things happened at this point, and we’ll never know which unless we can get the data from the black boxes.
SCENARIO ONE: they fail to take the proper actions in time and lose control of the airplane. This is pilot error, not software error, as the automatic systems are off line (baring the fly-by-wire system malfunctioning, but there is no reason to believe that at this time). If that happened, most likely they stalled the airplane, could not recover in time, and belly-flopped into the sea. As the airplane landed flat, based on the evidence, I’m inclined to think they stalled out of level flight and not in a turn. Level flight stalls you’re more likely to have both wings stall at the same time, resulting in a flat attitude. Stalling out of a turn you’re more likely to wind up in a spin, particularly if you’re having instrument problems, and that tends to go in nose-first at a high rate of speed (possibly exceeding Mach 1 in an airliner). Alternatively, a spin can develop into a flat spin, which would give you another belly-flop as the fuselage is more or less horizontal to the ground in that situation but I have no idea how inclined an Airbus is to get into that state, or how long or how much altitude might be required. Anyhow, they hit hard, and if they belly-flopped I’m not sure you could describe the mode of travel as “flying” any longer. That would be falling, and it’s very rare that airplanes fall out of the sky.
SCENARIO TWO: they do everything right, alter pitch and power for proper angle of attack/airspeed, and it’s not enough. A major storm of that intensity can, at least in theory, overpower an airplane (it can also tear it apart, but that did not happen in this case). In which case… they lose control, stall, and fall out of the sky.
It’s important to aviation to know which scenario is true, and preventing a future occurrence may require better pilot training (for one) or better airplanes (for two). It is possible that there were elements of both at work, which means improvements in both training and airplanes. Arguably, we should be able to get better weather information to pilots en route over the ocean, too. I’m sure someone is working on that, too. Clearly, avoiding these situations would be an excellent strategy.
Now, another point, and this one about icing. They talk a lot about icing on the pitot systems, and it’s true they can ice up. But supercooled water isn’t magnetically or magically attracted to just pitot tubes. It’s going to be hitting the entire airplane, and it’s going to stick on the airplane. This is bad for two reasons. First, it adds weight which is not a good thing for a flying aircraft. Second, and worse, it can change the shape of the wing. It’s the shape of the wing that provide lift. Ice sticking to the wing changes the shape, which can also change the angle of attack depending on how it builds up. That means, for an airplane accumulating ice, the margin between flying and stalling is reduced, and this is normally expressed as “the stall speed goes up”. What that means is that with ice on board the indicated air speed at which it stalls is higher than it would be otherwise at that pitch and power and everything else. You’re not generating lift as efficiently, and with added weight you need more lift to fly. If things get bad enough you can enter a scenario where even maximum power and the most efficient angle of attack won’t be enough to keep flying… and the airplane goes down. It’s pretty rare, as pilots try to avoid such circumstances. Also, in addition to pitot heat, airliners have other systems to avoid ice build up or get rid of ice from critical areas. Like pitot heat, though, these systems can be overwhelmed if conditions are severe enough. “Airframe ice”, as it’s called, may or may not be a factor here, it would depend on how much had accumulated. It’s possible to have the pitot tubes freeze over (as they noted in that list of prior pitot failures) and not get into further trouble. Airplanes can tolerate some ice sticking to them.
So, right now what we have is:
= Routine flight over the ocean
= SOMETHING BAD HAPPENS!
= ???
= Belly-flop
It’s the flight recorders that can replace the question marks. If we can pull the information off those units we might actually know what happened. If we can’t, we’ll never know.
