Apparently with enough power you can even fly without a wing (as happened to this F-15 after a midair collision).
I don’t know what NASA or similar research entities do with their own custom sims, and it’s possible that they use accident data in this manner, and I guess I would expect them to. I just don’t know anyone who I can badger with questions about it!
To make a new sim, though, one that will be used by airframers, airlines and training companies (CAE, FlightSafety) to train new pilots… this information isn’t used, at least not by the leading manufacturer of such full-flight simulators in the civil sector. When a brand-spanking new airplane is being developed, the experimental simulator is often developed alongside it or early in it’s commercial life, using theoretical data and data modelled from similar planes, ones of about the same size, engine type, etc (say, using Challenger 605 data as a preliminary engine model for the CRJ-1000, as a totally made-up example).
As the new plane goes through flight testing and certification, specific data is collected by the flight test crews - or by simulator company employees that go along for the ride - and that data is modelled via computer systems to run the environment in the simulator. There are standard bits of data that need to be collected under specific flight conditions which need to be replicated by the simulator in order for it to be qualified by Civil Aviation Authorities. This will include, for example, aerodynamic, performance and sound data up until stick shaker/pusher/buffet, but there is no requirement for data beyond that (for safety reasons) and accidents are not used to fill in those gaps. The FDR/CVR simply don’t record enough data in the necessary format to really use it much anyways - if an accident reveals something that was missing, the data will, if possible, be collected using an experimental test flight or will be theorized, but as I said, this is rare. Once the entire model is developed and the simulator is complete it goes through certification, and any further simulators are merely copies.
Updates to simulators tend to be fixing software bugs, adding capability for a new feature on the same plane (say, making an existing sim handle a new PFD unit or a HUD) or making a simulator “switchable” between two planes of the same “type”, like a CRJ-700 and CRJ-900.
Once again, any errors in this summary are mine though misinterpretation, not those of my friend who has been rather patient in explaining things to me over the years and since last night via text messaging! 
A couple of points -
As I understood the news release, they were flying through turbulence when things went to hell. The pilot put the nose up (or the trim did?) until the aircraft stalled. They held this attitude and stall until they crashed.
Sounds like (initial speculation was) they mistook stall turbulence for severe storm turbulence. One suggestion was they thought they were caught in a massive downdraft, plus wanted to climb out of the storm, so they nosed up. Without good data, they did not realize until well into the death drop (if then) that they were stalled and dropping, not in a severe downdraft. (Hence, “je ne comprends rien!”; they were still trying to get out of a downdraft.)
With the environment they were in and possibly incomplete and unreliable data, they probably did not understand what the heck was really happening. Yes there would be some ear-popping, but I imagine in a pressurized pressure-regulated cabin the effect would be minimal.
I recall in my pilot training, the instructor took us into a fluffy cloud (no turbulence) told me to fly straight by gut feel. After about 30 seconds, he pointed to the instruments - we were tilted and in a broad downward spiral at about 500fpm, but by “visual” in the fog it looked/felt straight and level.
If also had a situation in choppy turbulence where those “floating ball” instruments - level and compass both - were impossible to read within 30 degrees either way. I doubt non-electronic intruments would be terribly informative when you are being slapped back and forth 40 degrees, unless you stare at them for a while - which they would not have had time.
I used to get the Transport Canada bulletins for pilots on air accidents - it’s amazing how many “stupid” things people do to get into (and during) extreme situations, even simple things like “opps, forgot to put the gear down”. It’s easy to sit back with an armchair view and say “how could they not know?” but in reality, it’s an extreme situation and they were probably too busy just trying to regain straight and level to have enough time to process the bigger picture.
Is is accepted that the deep stall was induced by the autopilot, as opposed to control inputs made after the autopilot disengaged? And was the trim issue due to runaway trim, or to something the pilot(s) did?
Oh, absolutely - keep trying to solve the problem all the way down. After all, what do you have to lose, right?
A question of idle curiosity: What would the pilots have done if they had in fact managed to pull out of this seemingly harrowing stall? Would they have continued on to Paris? or would they have sought the nearest convenient runway? Would they fly at a lower altitude, perhaps?
It’s hard to say. I’d be avoiding icing conditions to start with, then whether I continued to the destination or not would depend on if I suspected the aircraft had been damaged. I think the most prudent option would be to go and land at the nearest airport that had suitable maintenance and passenger facilities.
Given they were over the Atlantic there really would have been no alternative to continuing the flight. I’m not sure what the closest facility would have been, although the Canary Islands seem a frequent destination for airplanes having trouble with the crossing.
Whether or not to continue to Paris is an interesting question. If they completed the flight at a lower altitude it would consume more fuel, so they might have no choice but to stop for refueling before reaching Paris. There might be concerns about damage to the aircraft. It’s hard to say. Presumably, after the immediate crisis was over, the pilots would have attempted to contact their dispatch, consulted with others, and come to some decision as to what would be the most prudent course of action.
Xema, you’re absolutely right to point out those two items.
Here’s the excerpt from the report discussing the initial event:
The way I read that, the autopilot and autothrust remained engaged until just before the airplane began rolling to the right, indicating a stall. The stall was not necessarily the autopilot’s fault, but in my book, when the AP disconnects and the airplane immediately stalls, that’s an autopilot induced stall. I have no idea what the angle of attack was at that time, nor what AoA would be considered a “deep stall” in the A330. It is likely that the stall was recoverable at that time, but the conflicting and wildly fluctuating airspeed indications combined with the severe weather would have made it much more difficult.
With regard to the trim, I had assumed it was activated by the airplane’s flight control computer(s). However, upon further review, I don’t know for a fact whether it was the airplane or the pilot that ran it up.
The trim may have run because the airplane was trying to keep the nose up as the pilot continued to make nose-up inputs. I don’t fly the Airbus, but I understand that under normal conditions when a pilot makes a nose-up input, the airplane will automatically trim to keep the nose at the requested pitch attitude. The airplane may have simply been trying to comply with the pilot’s request. Or, the pilot may have run the trim himself, the report doesn’t specify. The position of the trim switch should be recorded by the FDR, so we may eventually find out. There was some mention of alternate law by the crew prior to the third stall warning, which may or may not indicate behavior of flight control logic at that time. Again, I don’t know.
I wanted to briefly mention the continual nose-up inputs made by the pilot flying. Earlier in the thread we discussed the importance of lowering the nose during stall recovery. The stall recovery technique in the Airbus may very well be just the opposite, to hold the stick back and let the flight control logic reduce AoA and minimize altitude loss. I don’t know if that’s the case or not, but it’s a possibility.
I was a bit overzealous in crafting the conclusion of my first post, and I apologize. However, the fact remains that this was a sudden-onset maximum-altitude stall in a near-gross-weight airplane immediately following an unexpected autopilot disconnect at night over the ocean with conflicting flight instrument indications and simultaneous system failure in the middle of negotiating a large line of severe thunderstorms. That’s a nasty setup.
Could a false airspeed indication set off the stall warning alarm even if the plane was not in fact stalling?
Also, could a false airspeed indication cause the stall warning to NOT alarm when the plane was in fact stalling?
(I may as well pose my query, imbecilic though it may be.)
Assuming there was a minute to spare, should the captain have gone on the intercom and said:
“Ladies and gentleman, we’re experiencing a minor glitch with our trim controls. Would all passengers please stand up and walk toward the front of the plane.”
Would that have brought the nose down, at least if the situation weren’t yet completely hopeless?
I would rather die seated than packed among a bunch of sweating, shaking soon-to-be-dead passengers. 
If the plane was rolling violently and the nose pitched up a lot, they probably couldn’t have moved forward fast enough to make a difference.
Yes on both counts. In the 90s a 757 operated by AeroPeru took off out of Lima at night. The static ports were still taped over due to recent maintenance being performed (several people have to miss this for it to happen). The static ports are essential for proper operation of the pitot-static system.
Anyway, as they climbed out (over the ocean, at night) they got increasingly unreliable airspeed indications. At one point they had both the stall warning and overspeed warnings going off simultaneously. That was one noisy, confused flight deck.
It sounds like, from the data in that report, we don’t know for sure whether the plane was put into a stall condition by the autopilot.
Pretty sure we will never hear about it if it was. No way France will damage Airbus that badly by suggesting its planes’ computers kill people. People are edgy enough about flying without thinking Skynet will become conscious and crash them all into the ocean on a whim (hyperbole I know but then phobias are not really rational anyway).
If that is the problem it is not a likely one to occur again and Airbus will quietly make some routine software updates to fix it. Whether deserved or not I suspect the pilots will be taking the blame for this one.
ETA: Airbus will accept the slap on the wrist for the pitot tubes which they were already in the process of replacing anyway.
I agree, but from the sequence in which events are presented, the implication seems to be that it was the pilot’s nose-up control input.
The relevant paragraph:
I think we need to know how quickly that velocity change occurred. I interpreted it to mean that that’s when the speed readings failed, not that the actual speed fell to 60 knots. Also, I think when you bank left or right in normal conditions, the nose drops a bit and you pull up on the nose during the bank. At least that’s how it works in MS Flight Simulator – but I never flew an Airbus on it.
I don’t want to call any question stupid but let’s look at the facts. There were 228 people on board which meant the flight was mostly full. What happens people stand up at the end of the flight before the doors open in those narrow isles? There is no where to go right? The same is true here. An entire mass of people can’t just suddenly go to the front of the plane even if they had advance warning because there is no where for them to move to. Even at 170 lbs per passenger all inclusive, that is less than 39,000 lbs of self-carrying cargo that could move in a plane that weighs up to 514,000 lbs at takeoff. I can’t do the exact center of gravity calculation based on these types of assumptions but you see that the potential weight shift wouldn’t have had a great effect even if people laid on top of one another in unison. It may not have addressed the root cause of the issue either.
Sudden weight shifting by passengers has been done occasionally in much smaller planes as a joke to make the pilot have a less than elegant landing or so I have read in military stories but I don’t think it could work here.
Agreed - without the precise times that things occurred, it’s difficult to sort out cause and effect.