Proper stall reovery for JAR 25/FAR 25 machines
 

Guys,
Before they even got into the problems and confusion of the stall warner going off then on (below 60KTS) they ignored the stall warner sounding continously for 50 secs.

At no point was there any attempt to corectly deal with the problems eg PF " I Have Control - Action the ECAM items" no formal attempt to analyse what was going on - just an almost blind panic.

As has been discussed many many times on the tech threads when they reduced the power to flight idle the pitch decreased and the speed increased even with full aft THS. If Pitch down had been commanded the THS would have trimmed forward and together with the power reduction would eventually have resulted in a recovery.
The problem lies with inapropriate drills, a complete failure to follow the correct cockpit procedures and no command and control.
Reason - failures in the training they recieved and procedures they were trained to follow which is also a theme for the colgan crash.

Somethings that strike me now after this marathon.
A wry note.

The :oh: time taken from accident to report seems roughly normal for the French. :ok:

As usual no sense of proportion or anticipation of problems & adopting immediate solutions.
Instead it's left to the relatives of the deceased to continue to 2nd guess the "what if" scenerios just like when the survivors froze to death at Mt St Odile.

The tragedy has primarily been blamed on :bored: guess who?
(Same dead people)

Air frantic is clearly on the defensive, BUT....
big but (t)
...it now has one of the worst safety records in the world, and the price of rescuing the FDR etc has already been astronomic.
This company could face liquidation over this latest of many fiascos.

:ugh: If they continue to operate, who is going to bale them out this time?, or will they merely go for the standard French rebranding exercise for unpopular, unsuccessful or plain lethal companies eg*. ...

Let's not forget the current French government doesn't have any money.

Is it going to be hived onto the insurance industry or who is going to be behind coming clean for a change?

Agreed. After arriving at the correct control inputs through a process of elimination, the last thing that they need to hear was that they were doing it all wrong...

An experienced pilot would not have pulled up to 14 degrees at 35,000 ft and expected the aircraft to do anything but stall. He probably didn't know how the stall warning shut off below 60 knots either. An experienced pilot would have held cruise attitude around 2 degrees nose up, maintained cruise power and got a checklist every aircraft has called unreliable airspeed and saved hundreds of lives.

BUT THEY DIDN'T.

from avweb, prolly not news to a lot but however.The French news agency AFP reported Wednesday (one day ahead of the official final report's release) that investigators have concluded that pilot error and technical malfunctions caused the crash of Air France Flight 447 in 2009, killing all 228 aboard. "A source close to the case" told AFP that speed sensors on the Airbus A330 "froze up and failed" as the aircraft entered a line of thunderstorms while flying from Rio to Paris. That information was reported more than a year ago in factual findings. However, the source also told AFP that the official report of the French Accident Investigation Bureau, BEA, concludes that the captain then "failed in his duties," and "prevented the co-pilot from reacting.

So the sensors "froze up and failed" did they? So did the senseless ones up in the front.

When I go SLF again I hope I don't have such clueless losers piloting me. They should have started with breaking a few model airplanes. They might have learned the basics with small pain, and saved their profession the shame.

I had a quick read of the report conclusions and there seems to be an assumption that it is normal for the autopilot to drop out due to UAS. I can understand that the autothrottle requires airspeed reading, but a 3 axis autopilot should not require airspeed and does not on basic aircraft.

So if a Beach Barron suffered UAS, the autopilot would keep flying the plane giving the pilot a chance to work the problem.

Could not Airbus design a plane so that full control does not need to be passed back to the pilot should any input data including airspeed be incorrect? It would be easier for the crew to work the problem if partial automation was maintained.

Immediate actions should have been attitude and power. These instruments worked throughout. VSI and directional instruments worked throughout also. Not only did this guys ignore the basics, they ignored them for 4 minutes. I conclude that the basics were not taught properly, never mind the fact that they were given no opportunities to practise them. This is flying lesson 2 for Pete's sake.

The report doesnt emphasize the big problem that confused the crew in:

-The stall horn discconecting below 60 kts.
-The big problem the automatic Trim can cause.

The FBW in the airbus doesnt have artificial feedback on the sticks so an automatic trim becommes absolutely necessary for hands off flying ( elevator neutral to cg ) if we want to make flying in this planes bearable for the pilot in a day to day operation.
Therefore,this trim being full up if commanded to do so does create a problem after the crew still has not figured out they didnt need to pitch up so much... specially because the trim up will stay there unless commanded to go down... making a dynamic neutral stability situation WITH a pitch up and high AOA taking the aircraft into a stall.

Thats why there are laws in airbus aircraft...because if there where none, almost all pilots would stall the aircraft with that lousy philosphy and hence, protections are needed, but guess what... that day there werent any protections available after pitots got frozen.

And there in lies the problem, you can't expect the aircraft's automation to get you out of every problem. Sure, Airbus could make some changes, I would start with WoW component to stall warning, maybe even some way for both pilots to know what the other is doing with the controls. But if you sit in an airplane thinking that automation will save the day, then you have no business sitting forward of the re-enforced cockpit door.

There is a lot of basic airmanship missing from this accident. The "box" that needs fixing in this case is the one that is racked right between the two arm rests. It has double redundancy and is supposed to be the most powerful computer in the flight deck, the one that takes over when all the other ones fail. There is no way to replace that box, and any attempts at removing inputs into it can lead to this sort of sad and tragic result.

That is not just an Airbus problem, but one which the whole industry needs to deal with. Airlines being one of the main culprits with their accountant managing and rampant, unchecked cost cutting. Some things are too important to cut.

This is what happens when some short-term manager makes a decision he knows nothing about which has consequences years down the line, of which they where told about by professionals who new better at the time.

All stall related accidents could be completely eliminated by installing a $100 AOA sensor. In fact, why are we still even teaching stall speeds at all? AOA never changes with altitude, load factor, anything. It's the only thing we should be referencing.

Do you really think an AOA indicater would have helped these two when fully functioning attitude indicaters, altimiters, VSI's and GPS ground speed info didn't help them? I don't think anything but a fully functioning autopilot would have saved them. Sad, isn't it?

Just seems to me the report is biased against the pilots and pitots and is not emphasisizing all the design and instrument deficiencies of the aircraft. An AoA gauge should at least be there as a backup. I think the newer A380 has one.

In the confusion of all the other things going on, I agree an AOA would not have helped here probably as the system was based on airspeed.

But I think stalls should be de-coupled from speed completely. A stall should only be referenced as an angle. The speed tape in the MFD should be angles, not speed. Then we'd eliminate almost all stall related accidents.

If they had held their altitude the attitude indicator would be their AOA in level flight. Any experienced pilot would have done that. Unfortunately if you need an autopilot to do it for you, you are out of luck so you need to know how to do it manually.

OK, the BEA are saying that the crew on the day were part of the problem.

Maybe, or maybe the "system" was the real culprit, and a group that is never to be seen when things go pearshaped are more culpable.

How many long haul pilots on the Airbus or for that matter any other long haul aircraft have had any experience of hand flying at cruise altitude? Not many, and one of the reasons for that is the bean counters, who get upset that pilots flying the aircraft rather than the automation may cost more because it's not operating as efficiently.

So, most pilots have not hand flown in coffin corner during normal operations. Have you ever done that in the sim? Perhaps not, and that may be because cost benefit analysis done by (guess who) the beancounters, states that accidents such as AF447 are so rare, it's not cost effective to train to such high levels of skill. Really? I wonder if the bean counters ever travel by air themselves? One has to wonder. If I then ask how many people have had to deal with failures in the sim when hand flying at high level, it's probably an even smaller number.

How many of those beancounters and high level training people are actually at the airport comforting grieving relatives when the result of some of their policies is made painfully apparent? Not many, they are always well protected from the sharp end of such involvement.

Now let's go down another road. How many people have used devices such as a DVD recorder, and thought when using it, why did the programmers do the program in this order, it's crazy, the feature I use least is the first in the menu, and the one I need most is the last, and if you look into it, the reason is that the first item is probably the system set up, and nothing else will work if you haven't done that, and then they move on to other things, and the timer programming comes often last, probably because it's the most complex area, that needs all the others , but it's the routine that the user will use most often, so in a well designed system, it should be first on the list, and the lesser used items should be further down the list.

Same scenario is sometimes true in avionics, and in things like ECAM alerts, the order they are dealt with is sometimes a lot less than intuitive, and certainly not the order than a human flight engineer would bring then to the attention of the PF or PM, having taken them all into consideration and evaluated what they all mean. Computers are wonderful, but they have limits, one of which is that they only make decisions one at a time, and only in the order of programming, which in a lot of cases takes no notice or attention to the phase of flight or the state of other systems that may or may not have been evaluated in this iteration. Some of the ECAM type alerts that are thrown at an already busy crew are not critical to the underlying main and most important criterion. Aviate, Navigate, Communicate.

It is significant to me that in 3 incidents in recent time, the system warnings and alerts have in some respects got in the way rather than helping. The 380 that had the uncontained engine failure, they were a very long time working through all the check list issues before they could eventually put the thing back on the ground.

"Sully" Sullenberger commented that when he put the 320 into the Hudson, there "wasn't time to go through all the checklists", and that resulted in a vent being open below the "water line"..

The AF incident was made more difficult to deal with by misleading guidance on the instruments.

Then there's the whole issue of the stall, and awareness of it, both from the crew aspect, AND the aircraft, which was able to stall because the automation had already thrown in the towel and degraded it's operating mode, without making that very important fact known in no uncertain manner to the crew. Do you describe that as a design fault or a design feature.

Over 40 years, I've programmed many computers, and the "trick" to successful programming is to make sure that the right response comes out of the system even if some of the inputs are wrong. Getting the system, whatever it's meant to be doing, to reject input that's wrong, is more important than getting it to validate and work with the right data, and all too often, it's the manner in which the system deals with the errors that makes the difference between staying in the air or crashing.

There are other issues, like the 330 not being stable in pitch, which because it is flown by automation is acceptable to the people that certify it. Maybe, until as already mentioned, the automation throws in the towel. If as a result the crew is given an aircraft that's degraded, and also not stable, that's not exactly giving them too much assistance.

I'm not going to go down the road of the crew training and experience issues, I'm probably ruffling enough feathers and sensitive egos already, and to go there, and bring up issues like self funded type ratings, cruise pilots, and training systems that discourage people from using spare sim slots to improve their knowledge will only make the issues even more contentious.

Over 10 years ago, the EU funded a research project that was supposed to aid and improve the quality and capability of flight deck automation, by improving the evaluation and analysis capabilities of the systems. It seems that the changes that should have come out of that project are being stalled somewhere, as I've seen very little real change in the way that most of the systems provided automatically are operating.

I've seen this accident described in other places as Airbus' Titanic, and in many respects, it is, and how both Airbus and the industry responds to it will be very significant.

One thing is clear, for a vast number of reasons, many issues that have been raised in respect of training, experience, certification and design of flight deck systems are all going to have to be looked at in a very different light as a result of the findings of this report, the implications of this accident are as far reaching as Tenerife and Kegworth in terms of the issues raised and their consequences.

I just hope that all the people implicated are really listening.

In a strictly legalistic manner, yes, the crew of 447 on the day made mistakes that were contributory to the outcome. As to wether those mistakes were "pilot error" is very much a question that needs to be discussed in a lot more detail and over a wider group than even the inquiry team, and with wider issues than just the operation of AF447 in view. It seems to me that the 447 crew were as much victims as the rest of the unfortunate people that were on board the aircraft.

I am not sure that the BEA have succeeded in fully identifying the wider and larger issues that led to this event having the outcome that it did, or more specifically, I am not sure they have given enough emphasis to the underlying issues that are clearly implied in their findings. That may be because the media is taking soundbites, and maybe because of the manner in which their report has been presented. Hopefully, when the specific and detailed findings are analysed in depth, there will be changes at all levels of the industry,

I can only hope.

bubbers44:

I mostly agree. Except I would have placed the nose 5 or so degrees below the horizon, and reduced the power to "cruise" descent simply to go for better air, so to speak.

Errr, that would be Chesley Sullenberger. (when in doubt, Google). :ok:

Indeed, but that wasn't what they called him when they interviewed him on CBS news (via Sky) while I was typing my original. His comments were interesting too, they were in a Sim at the time, discussing the BEA report, and the crew response to the problems, which he found "strange"

Let's be clear. There was pilot error. That is a fact that is indisputable among fair and rational people. The pilots neither created the correct mental model of the incident nor implemented any of the correct procedures to respond to reality effectively.

How much the crew is to blame for their errors is a different question. I certainly agree that a technological failure put them into a bad position and that technology could have done a better job of helping them out of it. At the same time, insofar as the human crew is the ultimate back-up to the technology the crew failed in their duty.

I'll leave the precise apportioning of blame to others. Suffice it to say that there is plenty of blame to go around and the crew cannot be exempt.