roulishollandais

@henra

Conditions of "validity" are unclear, here, what is "valid" (two ADR and two IR FMGEC), when is it "valid", aso.
Which max difference is acceptable from one ADR to another ADR, from on IR to another IR, aso. A33Zab's shematic is unclear too, "compare" is unprecise, as there are difference which are acceptable, and in + or in -, so the shematic is much more complicated

syseng68k

infrequentflyer789, #288

Agreed, but in the existing system, pitot data will be subject to filtering
and may already include historical data to provide some sort of predictive
capability. The processing will reject out of range transient data. When a
storm is entered, things get more dynamic, but again the software will adapt
to this, changing it's filter parameters if required, to optimise the
measurements. If this were not already the case, flight deck values would be
unreadable.

It depends on how the data is processed and it's not a case of one data source
or the other in competition, rather a synthesis of the two, where the error
characteristics of both are given less weight in the calculations and the
strengths given more weight. The idea is to model the characteristics of all
the sources under all conditions of operation. The filters and processing can
then be matched to provide a more accurate and consistent result than from either
source alone. A more robust system in terms of transient events and redundancy...

syseng68k

dClbydalpha, #292

Sorry, missed this:

No idea, but have seen examples of gps in lrg ins units
(Ferranti) that date back to the early nineties, so it's not exactly a
new idea. If it's not used in civil aviation, it's not because of the lack
of well proven technology.

The first result from plugging:

"airbus laser based air speed measurement" into google:

Airspeed measured by laser | NRL Annual Report 2009

It will come, not a matter of if, but when :-)...

roulishollandais

@syseng68k
HN39, IF789 seem to have the same frustrations as iself with the final report's unsuffisant precision : all that is litterature, "communication" for commercial use, not science, not proofs...

mm43

A March 2012 comprehensive update on the NLR Daniela Project - a PDF file.

Lonewolf_50

Not completely pointless.

I posit for you three different crews, A, B, and C.

I sit as master torturer running the sim, and proceed to add task loading (via failures and malfuncitons) to the point where the crews fail. As we have discovered, we can usually find a way to task saturate nearly anyone, and nearly any crew.

For the sake of illustration:
After cascading seven failure tasks, Crew A are done for.
Cascading of 9 failure tasks does in Crew B.
Cascading of 13 failure task finally does in Crew C, and the sim torture master only had one more trick up his sleeve.

As each crew ultimately fails at a different level, I as training department, or as chief pilot, or whomever, have some information upon which to help inform who needs work on what, in terms of the training and proficiency requirements and systems knowledge.

Likewise, the crews may, after giving me the occasional dirty look, be aware on their own of some limits or holes in their knowledge of systems that they were not previously aware of. <== That is crucial to most pilots I have ever met, all of whom like to be good at what they do and tend to correct or improve areas where they find themselves not up to scratch.

For What It's Worth, not without value.
In the commercial world, however, there is a fine decision to be made due to cost of sim time, in terms of how often and how brutal, and what to emphasize. There are not infinite training funds.

bubbers44

Dozy, I agree with what you are saying. If they hadn't pulled the nose up to 15 degrees none of this would have happened.

Software isn't going to prevent a repeat of this event, competent pilots who know how to fly without an autopilot will. Unfortunately this breed is going away and Airbus convinced everybody that stall recovery is not required in training because the bus won't stall in normal law.

UAS checklists are in all airliners and would have prevented this crash if used. The PF thought he was protected when he pulled back the SS. They were not in normal law so they stalled. Both pilots in the cockpit needed the autopilot to fly with no airspeed.

DozyWannabe

Well, hold on just a minute. Airbus said you can't stall the aircraft in Normal Law - but it doesn't follow that such a statement should convince anyone that stalls won't happen. To my mind, airlines didn't ask the important question of what happens if the aircraft *isn't* in Normal Law before making decisions on training. The same thing happened 40 years ago when Douglas assured the industry that losing all three hydraulic systems on the DC-10 was impossible. Thankfully the captain on the flight where Douglas was proved wrong didn't buy it for a second and used his seniority to train himself how to fly the thing on asymmetric thrust.

With all due respect, that's conjecture - there's not enough evidence to say why he pulled up the way he did.

Airbus have had their problems over the years, for sure - but to say they're largely responsible for the decline in hand-flying skills is like arbitrarily blaming Ford for the fact that so few Americans can drive a stickshift - the industry and customers moved in a certain direction, and Airbus (and Boeing, MD etc.) supplied products to fit that demand.

The truth is that with ever more crowded skies, automation is going to be more necessary than it has been in the past (with the advent of GPS and RVSM airspace) in order to maintain a degree of safety in traffic. This is a distinct issue from the lack of training for situations where the system fails however - PJ2 did a rapier-sharp dissection of industry attitudes on the previous thread, but what it boils down to is the trend whereby more and more senior airline management have never developed specialist skills relating to the industry, instead being purely business-orientated.

Lyman

Dozy

Your "15" degrees PITCH UP I believe came from the DFDR, via ins and archive. You know the Pitch attained was not nearly that high, in the initial command post a/p loss. Once maneuvered, there ceased to be a comfortable baseline, and his PITCH was not available via INS, and you have no idea whatever what the cockpit was like, you just don't....No PITCH was ever spoken in the CVR, only "you go up....so go down..." "Am I in climb?" So your reference to a specific PITCH value from the DFDR is completely "out of your seat cushion"....I submit PF's grok of PITCH was never accurate, and from the evidence, there is NO way to know what they saw, or believed......via their instruments....

DozyWannabe

He yanked the sidestick halfway back the second AP disconnected Lyman - there's no getting round that.

Yeah, the instruments were clearly lying or absent and the aircraft tricked him into crashing. Truly, denial is not just a river in Egypt and I don't see the point in continuing this debate.

john_tullamarine

Re this post .. sideline philosophical consideration ..

As each crew ultimately fails at a different level, I .. have some information upon which to help inform who needs work on what....

Likewise, the crews may .. be aware on their own of some limits or holes in their knowledge of systems that they were not previously aware of.

My original comment was directed at routine sim work and I stand by the philosophy for routine (training value emphasis with a lower case "c" checking strategy) sim work.

In the specific circumstances you cite, I am with you .. indeed, I would be in the line to participate for the reason you consider .. just where is my personal brick wall ? The underlying research value for training strategies is valid .. but the participants must be willing and not at jeopardy.

Perhaps we should have a beer or ten to debate the pros and cons ?

OK465

727 2-engine out, manual reversion, NDB approach to NDB minimums was essentially a one maneuver check-ride.

What was most interesting, was not who succeeded or failed at the maneuver, but who would be enthusiastically willing to attempt it.

bubbers44

With all due respect, that's conjecture - there's not enough evidence to say why he pulled up the way he did.

Then why did he pull up into an impossible climb that all of us know would end up in a stall? No airliner can go into a 15 degree climb at FL350 and not stall. He either trusted Airbus protections in normal law or was incredibly stupid. I think somewhere in his training he was told you can't stall in normal law so pull it up and let the automation handle it. These two spent most of their careers monitoring autopilots so probably depended on them a lot more than us old timers to fly the airplane.

The cockpit voice recorder shows how confused both pilots were when AS and AP went away. It didn't happen like that years ago before full automation took over.

DozyWannabe

That's two possible conclusions. Other conjecture has included the ideas that he tried to chase the intermittent FD bars and/or that the pull up was a panic reaction - and through a possible combination of body-clock offset, having just come off of holiday (and still getting used to being back in the saddle) - or even just that the shock completely threw him, all training went out the window.

Unfortunately there'll never be any irrefutable proof which of these theories, or any combination of them, caused him to react the way he did.

I've said this before, but it's very important to draw a distinction between FBW/protections and automation (in the autoflight sense). The ColganAir Q400 did not have FBW or protections and yet the captain still pulled the aircraft into a stall when the warning startled him.

The distinction is important because (for better or worse) a pilot expects to be using the automation on almost every flight, whereas in an ideal world the pilot would never have to rely on the protections (and in the real world, most never have). The protections are there primarily to help the pilot stay within the load limits of the airframe when aggressive maneouvres are required - but in this case there was no requirement for an aggressive maneouvre and hence no need to bring the protections into play.

The "old timers" didn't have to deal with RVSM airspace.

More seriously though, the industry needs to have a serious look at maintaining basic aeronautical knowledge, now that automation has become a necessity.

It did on Dynasty 006 in 1985 and Eastern 401 way back in 1972. We need to be careful when generalising.

bubbers44

Yes, pilots have made stupid mistakes in the past and will in the future. We have made wonderful improvements in GPWS and TCAS so flight safety is better than ever before. We just need to not let the pilots become monitors of autopilots instead of pilots in my opinion.

In my career I had to continually improve my basic skills because the airline didn't. You don't have to rent a Cessna, just disconect the autopilot and fly on raw instuments when it is appropriate. Why would you want your basic piloting skills going away because the airline you fly for doesn't care?

You should care.

DozyWannabe

The problem is that the number of appropriate points where this is possible is dwindling.

I do - otherwise I wouldn't be here!

jcjeant

This is the kind of recurring comment
I wonder why fly in manual seems to be a problem for pilots who always invoke some of their company regulations
In fact is that these regulations have the force of law ?
Is there a law that prohibits pilots fly manually ?
Is that pilot can legally be put out of work for this action?

Machinbird

Unfortunately, some of the large but unenlightened airlines do hold to a punitive view of what pilots should do and should not. I recall a recent post in the Rumors section where a Captain was quite certain that he would receive an invitation to "tea and dates" with the chief pilot should he choose to practice his manual flying skills outside the narrowly constrained takeoff and landing portions of the flight. Apparently the statisticians at his airline had built some sort of case that to do so increased the risks.

To my mind, this is the short term risk avoidance getting out of balance with the long term consequences of such a strategy. Of course, if they bring in highly trained crew continually into their operation and flush them out of the system after ~4 years or so, then the consequences of their short sightedness could well fall to someone else.

Lyman

Hi Mach.

From above... "Apparently the statisticians at his airline had built some sort of case that to do so increased the risks."

The statistician is correct of course, at any given time, manual flight is more risky than autoflight. Clearly, the case can be made that it is "Training", a more demanding regime than auto cruise. In my view, the ball is in the regulator's court, and it would take a pro-active reg to cause the lines to upgrade hand skills.

Besides nil, zip, and no chance, what do you think?

slats11

True. But pilot and aircraft are inter-dependent. The aircraft design is clearly to revert to the pilot in the event of a problem.

An inherent assumption in this design is that the pilot is capable to manually fly the aircraft - without warning, at cruise altitudes, and with degraded systems or data. Increasingly, this assumption may not be entirely valid. What training is required by the operator? What is covered in the sim? Is regular maintenance of manual flying skills permitted by the operator?

Truth is often the first casualty of war. And it may be hear with the lawyers circling and large sums at stake.

However we need to ask ourselves if something has fallen into a crack somewhere between the designer / manufacturer, and the pilots. If there is such a crack, it is only a matter of time before someone else falls into it.