As I understood switching the AP on was not an option, it wouldn't help much.
Yes. If your airspeed is unreliable, AP engagement won't work. (Anyone, please correct me if I said that incorrectly).
I know that pilots were not aware that the indicators are OK, that was the "catch". My way of thinking was: "Nothing helps, we are falling, let's try with the AP switching on." Since I am not a pilot, that way of thinking is probably wrong
Yes, that is wrong.
Fly the plane, engage the AP once you have the plane behaving as your mission profile calls for.
I watched the video on youtube where one investigator (I assume) said something like: "Most of the people were probably sleeping, so they did not know what was happening".
That is a guess, but a good one.
In ten threads of discussion here, over three and a half years, a good analysis shows that the plane was in a modest nose up attitude and modest bank, for much of the descent and at impact. You'd want to read the FDR traces to see where it was nose high, lower, and how often the roll reached peak values. The nose was somewhere around 11-15 degres nose up when it stalled. Not quite as high at impact.
It seems that the aircraft did not have a significant airframe buffet while stalled. Following that line of thought, you would not "feel" it as with some other planes that shudder a bit more at stall, and when stalled.
From that (and my experience with traveling by planes) I concluded that passengers felt they were falling from the sky. I felt every time when the plane was decaying due to turbulences. Of course, this issue is not relevant to the air crash, just a little bit of human curiosity.
Probably not.
Once the plane settled into its descent, the "feel" would return more or less to "one G." The feeling you refer to is due to a transient condition. To apply what you feel as a passenger, you would expect the that the "feel" would be in the transitions, when something changes:
first, the climb before stall,
and then the onset of descent as the stall began.
Once established in the stall, the plane was more or less in a steady state descent.
You said that it was irrelevant if the indicators worked fine, because the plane was stalling. The point is that obviously they did not belive that indicators are OK, because in that case they would believe that they are stalling and they would have done something about it.
No.
They knew the indicators were giving them trouble, but you can't automatically infer from a bad indicator that you are stalling from that point of info alone.
A sound instrument scan should have alerted at least one of the pilots of the initial climb. At one point it did. He tried to get the other to "go down." The increasing rate of descent as the stall was manifested and began to take effect ought to have alerted one of the pilots, except for the point made that actual stall recovery on instruments isn't typically trained for. Stall prevention is.
Altimeters and Rate of Descent displays were both indicating a falling aircraft after the stall.
As I understood,
1. although there was a lack of training with high altitude
stalling,
2. the point is that pilots DID NOT BELIEVE they were in the stall.
3. If they belived it, they would have done something about it.
1. This appears to be a contentious point, since the industry has had upset training and understanding of upset for some years. Since certification didn't require exploring the stall envelope, the training is mostly in stall prevention, not stall recovery.
This does not appear to be an issue only at Air France.
That said, if you prevent stall, you don't need to recover from one.
2. Not "did not believe" they were in a stall but "did not recognize" they were in a stall. It can be reasonably concluded that what
they didn't believe was the audio
warning of stall. But we are indulging in a little mind reading, since not everything they thought is captured by something said on the CVR. Be careful.
3. I agree with this point, others may not.
2. The pilots did not understand what was happening because they did not belive to the instruments.
This is a debatable point. I would not put it that way.
What appears to have happened is their instrument scans broke down, to one degree or another, and the most useful information on the instruments was not what they paid the most attention to.
The AP was off, so they belive the stall alarm was also false.
No. That does not follow. Why they did not address the stall alarm, or comment on it, isn't so simple. AP being off had llittle to do with that.
My question here is: Is there any chance in that conditions for pilots to figure out what was happening, just relying on the cockpit view and the feeling?
No. They were flying in instrument conditions at night over the open ocean. You don't fly by feel in that case, you fly using you instrument scan. At least, you do if you don't want to get fooled by feeling and die.
That is a lesson written in blood over decades of flying.
I read that in stall conditions there is some kind of "buffeting".
That is a debatable point for the A330. It appears that the A330 doesn't buffet as much while stalled as some other planes do. gums described this as a "mushy" departure from controlled flight. This provisional conclusion was based on some detailed anlaysis of the FDR by some posters here, and I ask you to search that yourself.
I did not pay attention to that captain's "It is impossible". So that was probably related to the stall warning as you said.
Or not. There was so much going wrong that we cannot be sure what evoked that response.
Another thing. From the last 1/2 minutes or so I conclude that Robert and Dubois realised the plane was stalling, after Bonin reveales that he was holding the stick back for the whole time? After that they were "mad" and try to put the nose down, as I understand.
I'll leave that for others to guess at.
The other two may have figured out that they were stalled, but they didn't verbalize that based on what the CVR info we have available. Someone did a great job of reading out the decreasing altitude in the end game, per the CVR, but that didn't translate into effective action to recover from stall.
We have little evidence of an attempt to recover from stall by lowering the nose (some of the more informed estimates arrived at by various pilots here are that recovery would have had to begin somewhere between 12,000 and 20,000 feet to get unstalled and flying again. There were a number of discussions on how to figure that out, and I don't think a single conclusion was agreed.)
I don't recall who, but somebody did try this in a simulator. (Caution: the sims need data points from real flying to create good simulations, and nobody has done test flights of this plane into and recover from, stall.)
Based on the CVR and FDR, there is some evidence that, without saying anything about a stall, there was an advance of the throttles to TOGA that
might have been related to the sound of stall warning. Granted, this was probably based on a drilled response for a problem at lower altitude, but once again there is a bit of guesswork involved here.
Look up Unreliable Airspeed procedure for more detail. Lots of discussion on that. (UAS is I think the acronym used in these discussions).
As I noted above, if you aren't already stalled, added dpower might help, but if you are already stalled, that may hurt your chances of recovery.
Do you think that is likely?
A lot of things are likely.
What was unlikely was that AF447 fell over 35000 feet to the sea from straight and level cruise flight. But it happened anyway.
You may wish to look into the discussions of Thales versus Goodrich pitot tubes, or pitot probes, to understand why the airspeeds became unreliable and got this chain of events started.