Old Carthusian

Machinebird
Sorry for the delay in responding but I would submit that what you were writing about is the perception not the actuality. Some commentators are adopting a them and us approach - pilots and airline companies. It more suggests to me a misunderstanding of just what it takes to run an airline than an actual perception of reality. It also ties in with a common human perception that things were better in the past whether it is true or not. It may indeed be that some things were better in the past but it is also true that some things weren't.
When you actually study some of the older accidents (and you don't even need to go back to the 70s to see this) you realise that there have always been pilots who have lacked the necessary skills. What automation has done is allowed a massive increase in the number of flights with a concomitant increase in the level of safety needed to sustain this. Of course you will find some negatives in the introduction of automation and the way it is used but this is due to the humans involved NOT the machines. It is clear that the skills of this particular crew played the significant role in the accident but this is not a general malaise as this accident is unique and other crews have dealt successfully with similar incidents. This alone should be sufficient to demonstrate that the accident is not indicative of a general problem.

jcjeant

Hi,

Maybe not a general problem ..
Not related with this particular accident .... but the fact is that for decades .. automatism or not .. the human part of the plane does not seem to increase in security ... it remains stagnant in the statistics .. even with all the progress
Accident statistics
1001 Crash - Airplane accidents statistics - Crashes causes and aeronautical terrorism

Lyman

I think it is a reach to even call this tragedy an accident.

If all standards, precautions and minima had been met, this wreck does not happen. End of.

The population of UAS incidents was low, 32 I think, 6 involving AF. Including them in 2 Sigma is a breach of courtesy. The very availability of Back up Speed Scale on Airbus suggests an awareness of the problem, plus the AD on Thales, and the mandated r/r. The threatened RTF by the pilots shows they thought along those lines.

This crash doesn't belong in the gross numbers, it skews the reality like a bookie having a bad and "middled" day.

Flying is far safer than this c/f suggests.

Mr Optimistic

PJ2, thanks for your time - it's a pleasure.

Old Carthusian

Jcjeant
But as a function of the number of flights - human error is decreasing. The number of HE accidents remains fairly consistent but the number of flights have grown significantly. Partially automation but also partially better standards and culture.

bubbers44

Pilots need to know how to hand fly when the autopilot disengages. Usually just holding present attitude and power will handle it while figuring out the details. We have known for decades pulling back at 35,000 ft will stall any airplane so why do it? If you can't remember what your attitude and power settings were just set attitude about 2.5 degrees up and power to where it normally is until you figure out what you want to do.

I have no idea why these guys did what they did but no competent pilot would do it.

PJ2

Turbine D;

No worries, I understood what you were explaining and am somewhat familiar with Six Sigma - it was actually quite a good explanation, thank you.

In response to your important comment: Yes, I think so - we're thinking along the same lines and in fact I worked on a design change of this drill last summer, (July, 2011). I have always considered the Unreliable Airspeed memorized drill poorly designed and confusing primarily for those circumstances in which I considered the safety of the flight "not impacted". Here is why:

The UAS memorized drill has a primary decision point, "If the Safe Conduct of the Flight is Impacted", which requires that the crew decide "are we safe or unsafe?" In my view, this primary decision point which is intended to lead the crew into one drill or another, is too imprecise. The "one way or the other" is either to pitch up to certain degrees and set power to TOGA or CLB, or to not execute these memory items and instead "Level Off > Troubleshoot > QRH Pitch and Power tables". In other words, don't pitch up. To my knowledge this crew never experienced this kind or level of training in their recent sim sessions, (and I can categorically say that I have never even encountered the UAS drill in all my Airbus A320/A330/A340 initial or recurrent ground or sim training.)

In a recent and much more detailed description of this drill than I have seen before, a note in the PPT indicates that the meaning of safe conduct will be defined in training. Then it goes on to explain how the drill is done. The presentation is dated 2006.

In other words, "when" the safety of the flight "is/is not" impacted has a definite meaning but this is not defined in any SOP/QRH/FCTM with which I am familiar.

But this isn't the important point. The important point is, no matter what phase of flight, the primary "bifurcation" point...the point where the memorized drill launches one way or the other, applies to all altitudes and all phases of flight - climb, cruise, descent...and, as you have suggested...there is no defining of "cruise", or "above FL200", etc etc.

So I thought why not base the primary "bifurcation" or decision point on phase of flight rather than the very subjective and individual assessment, which defines the "safe conduct of the flight"?

Clearly, the safe conduct of the flight is impacted at/during the takeoff phase...the closer to the ground, the higher the risk. This is a direct result of the two loss of airspeed & altitude information accidents we are now familiar with, the Birgenair and Aeroperu B757 accidents, (which, for others, please see).

So immediate pitch and power numbers need to be memorized and instantly applied until "at/above MSA or Circuit Altitude" and then the aircraft is to be leveled off for troubleshooting. Leveling off includeds selecting the GPS Altitude on the FMC/MCDU, and getting out the QRH pitch and power tables.

In the memorized drill, there are three sub-decision points depending upon where the airplane is in the takeoff/early climbout phase.

For information purposes, the Airbus thrust reduction altitude in force at the time was 1500ft AGL. I believe it has since been changed to 1000ft AGL, but no matter.

The three conditions of the memorized portion of the drill (and after the decision is made that the "safe conduct of the flight is impacted") state:

Pitch/Thrust:
- Below THRUST REDuction ALTitude..................15°/TOGA
- Above THRUST RED ALT and Below FL100........10°/CLB
- Above THRUST RED ALT and Above FL100.........5°/CLB

The FCTMs I've seen further confuse the issue by saying first of all, if you've chosen 5deg of pitch because you're above FL100, (thinking the "safe conduct of the flight is impacted"), then very quickly get out the tables to ensure speed stability, (specifically, the overspeed sitution is mentioned...nothing is said of the opposite problem...loss of speed if pitched up too much). But then the FCTMs state that if the safe conduct of the flight is NOT impacted, the crew will not apply the memorized items but get out the QRH pitch and power tables while maintaining stable flight.

So the potential for applying the wrong procedure is there, but the potential is largely molified by the actions described in the QRH and as such reduce risk providing the crew gets out the tables quickly and establishes stable, level flight.

The difficulty comes when the subjective assessment of the "safe conduct" of flight is made...some would consider the loss of all speed indications a clear and present danger to the flight and launch into the drill. Others would light up a pipe, (metaphorically speaking...to indicate a calm, measured approach! ) and maintain level flight while the other pilot got out the numbers. Why not base the decision for subsequent actions on the phase of flight, where there is an emergency, and where there probably isn't, (as in cruise flight)?

It made sense to divide the phases simply...into "Takeoff or Below MSA" and "Above MSA in Climb/Cruise/Descent.

For the former, to the usual memorized items. For the latter, level off for troubleshooting, set the GPS Altitude and groundspeed on the MCDUs and get out the QRH pitch and thrust tables.

Machinbird

OC, the delay was expected due to time zones.
There has been a major de-skilling of the airline pilot community by virtue of automation. The new guys coming up are very good in handling the automation and there are apparently an increasing number of them who when asked to hand fly an aircraft break out into a cold sweat. The environment makes it difficult to acquire and maintain essential hand flying skills. The periodic simulator training sessions are too infrequent to really maintain hand flying skills. Many of the formerly accomplished hand flyers have commented on their personal loss of the touch. There is no doubt that automation has permitted a high level of safety despite this apparent loss of skills, but when an aircraft loses critical systems and the automation is crippled, are these new pilots ready to take over and fly?

When I was actively flying, such a loss of control as AF447 experienced for the reason it lost control would be unthinkable. The weakest pilot in my squadron could fly solid instruments by hand (Where we sometimes saw problems was in headwork.)
With 32 UAS events and 1 loss of aircraft, the statistics for that condition are terrible. Is it a statistical fluke? I don't think so.Statistics doesn't work like that. UAS is clearly much more hazardous than ordinary flight.

IMHO AF447 could well represent the "canary in the coal mine" warning us that the hand flying deterioration has begun to cripple not only the third world airlines but also the legacy carriers.

Any airline pilot should be able to fly cruise by hand, cold without a warmup. If he cannot do that simple task, then he really doesn't belong in the cockpit.

Old Carthusian

Machinbird
I am still not convinced by this deskilling you write about. So far what seems to have been discussed is a perception that this is the case which may be widely shared but is just that - a perception. A lot of people saying something is so does not necessarily make it so and given that accidents due to pilot error have remained fairly constant with the massive increase in flights across the globe the statistics would not seem to back this up. I have no doubt that there are pilots who react as you say to the challenge of hand flying but whether this is a general case I am less sure.
I suppose what I am saying is that one must be careful not to over-generalise and to fall into the human error of thinking that things were better in our younger days. Military pilots were notorious (and probably still are) for their high crash rates but this is the price you pay for pushing the envelope as it were. The type of accident represented by AF447 may well have been impossible in your squadron but possibly your colleagues found different ways to have accidents? Error does not manifest itself in a consistent pattern. Where I do think you have a point is that 32 UAS incidents do indicate an issue that does need attention.

gums

'bird's last comment is also what scares me.

I fully realize that many of the heavy pilots here have not flown to the edge of the "envelope", or exceeded it. I have no problem with that.

What scares me is a basic lack of airmanship I would expect of a teenager that I was helping learn to fly in a Cessna.

Lost a friend at Cali back in 95 or 96 or...... Stoopid flight management system turned the jet the wrong way and they noticed the error but kept descending whle turing back to the approach fix. Not good.

The increased use of automation seems to be a significant factor in recent incidents.

I would prefer a crew that flies the plane using a combination of "auto" aids and manual kills. IMHO, not enough manual flying these days. I am not a dinosaur. Flew with the latest and greatest avionics and FBW and such since the 70's. I would simply hope that the folks up front in the cockpit can actually fly the plane when all the automation and computer stuff goes away. In short, I want a "pilot" up front.

HazelNuts39

PJ2's sim pitched down with "full forward stick applied and held (...) from an AoA of 40deg to 10deg (in) 24 seconds", i.e. at an agonizingly slow rate of 1.25 deg/second. How are you going to improve on that? Using manual trim might help, but see Dozy's post #1096. Pulling the thrust levers back to idle would help to increase the ND pitching moment, but that's about all I can think of.

In this initial phase of recovery the sim is performing outside the envelope of known aerodynamic data from flight test or windtunnel. It is possible that actual airplane performs better. On page 111, at 02:12:45 in TOGA and with the THS at -13.8 deg, the airplane pitches down in response to a short push on the stick at 2.7 deg/second, and again at 02:13:30 at CLB thrust.

rudderrudderrat

I agree.
They can in Normal Law - no problem.

On a non FBW aircraft, in the cruise we'd take the autopilot out occasionally, enjoy the feel of the aircraft's response, re-trim it then re-engage the AP.
On AI FBW, there is no requirement to re-trim - it's done automatically. Even if you take the AP out - it's still in Normal Law and handles differently to ALT Law.

There is never an opportunity to practice Alt Law flying skills.
I think the lack of opportunity to practice manual flying skills in Alt Law needs to be addressed in recurrent sim training opportunities..

PJ2

HN39;

Re, "Pulling the thrust levers back to idle would help to increase the ND pitching moment, but that's about all I can think of."

The TLs were in the CLB detent in the exercise cited, so yes, that would have helped in the speed with which the AoA reduced.

And yes, recovery from AoA's of around 16° (just before the apogee but well into the stall warning) was a lot quicker - between 10k and 15k with more altitude required for the pull-out once unstalled, (secondary stall occurred a few times when we pulled too hard). The exercise described was a recovery from a fully developed stall, (AoA > 40°).

RR_NDB

Hi,

gums and OK465

AA flt 965 almost cleared the hill by crew action (based on GPWS). Crew errors, data base deficiencies, etc. played a role.

Clever System was developed and introduced "in the aftermath" of this accident.


PS

The combination of errors was reflected in the percentages:

In June 2000, the jury found that Jeppesen was 30 percent at fault for the crash, Honeywell was 10 percent at fault, and American Airlines was 60 percent at fault.

PJ2

Hi gums;

Re, "Stoopid flight management system turned the jet the wrong way "

IIRC, towards the first selection in the NDB list, "R" which was way east, and not the approach "R" (Rozo) NDB to which they had been cleared. In the B767/B757 one has to execute the selection while in the Airbus FMC once selected the airplane is going to go there. I always thought that was not a good design precisely because of this accident but it's still that way.

Painful I know - I recall the Dec 20 '95 accident well and know you lost a friend. You can take some solace in the many positive outcomes regarding CRM, SOPs, automation behaviours/procedures but especially Don Bateman's/Honeywell's work on the improved EGPWS system.

On the recovery altitudes, I think familiarity with the boundaries of controlled flight, which comes with an understanding of aerodynamics of your airplane would help but of course that's not what we do nor are we taught a high level of aerodynamics regarding our designs, and, we have no business being there at the boundaries! Not even test pilots actually stall the aircraft anymore as you know.

The exercise was a worst-case - fully developed stall, late recovery attempt, thrust was not idle. As I mention to HN39, a smart recovery, (as in brisk forward stick at the first stall warning blip, held fully forward without variation, thrust at idle), can be made which reduces the altitude required. It's still going to take a lot of altitude to a) regain the AoA and b) regain the energy, (due low availability of excess thrust, so its height for energy, initially).

Lyman

"In the B767/B757 one has to execute the selection while in the Airbus FMC once selected the airplane is going to go there."

Not always, though I'm sure statistically reliable. I have heard (anecdotally) of the bus abandoning the Flight Path in favor of a turn, caught by pilots who then fly manually to destination. Since I am unwilling to be more specific, note I use the word "anecdotally". These occurrences, upon follow up, were not satisfactorally explained.

Have you heard of this?

On pp9-10 of BEA #3, there is a flight path narrative. On pp29-30-31, graphs indexed to CVR. The latter is more explicit, and one assumes, more reliable.

On the last of these three pages, at the top "Thrust levers to TOGA" (DFDR), there is an interesting flow of data, conversation. With TOGA, the RHS inputs 3/4 stop NoseDown. Strange, eh? The LHS, just prior to RHS "I no longer have control of the a/c" states: "We have the engines....what's happening?" If he was concerned about AoA/STALL, why would he take note of the Thrust at FULL as somehow comforting, instead of a dire threat to LIFT?

So it is strange to entertain that the pilots (including Captain) would be concerned with STALL (high altitude), and be OK with the engines at 100%+ N1?

At the very least, why did BEA choose to truncate the CVR just as the controls went LEFT, the PF had lost control of the ship, and the Captain enters "What are you doing?"

Is it a tease? Obviously there is a reason. What is in there that cannot be seen by the public?

PJ2

Lyman;

Re, "I have heard (anecdotally) of the bus abandoning the Flight Path in favor of a turn, caught by pilots who then fly manually to destination."

"Flight Path" is defined as the vertical element while "turn" is obviously the lateral element. VNAV requires LNAV because the descent is defined in terms of altitude and possibly speed constraints which are either part of the STAR arrival or merely slow-down or speed control points such as found at Heathrow or Frankfurt. Put another way, the programmed track between waypoints is necessary before the vertical element (VNAV) can manage the descent to achieve the required constraints.

If radar-vectored off the course programmed, the lateral/vertical elements are no longer managed and become instead selected, (short-term operation), and HDG is selected, the lateral element is removed and the vertical element is logically no longer capable of managing VNAV calculations due absence of waypoint constraints, therefore the vertical reverts to Vertical Speed, or more rarely, FPA, neither of which are 100% suitable but for unimportant reasons. The PF can then choose a more suitable mode such as Open Descent - the Alt Sel on the MCP is reset to the cleared altitude if it was previously set to the lowest altitude on the STAR...some airlines allow this, some don't, "just in case."

The Airbus doesn't abandon anything even if in VNAV it is going to miss a constraint - in such a case it will signal on the PFD the need for speedbrakes.

In short, the scenario you're describing can't logically occur and there are no Tech Bulletins extant describing such odd behaviour so I doubt it occurred at all. Leaving the flight path does leave the FMC-programmed LNAV course and the autoflight lateral mode drops to HDG but that is completely normal behaviour for all these autoflight systems, with minor variations on the theme. If what is meant by "fly manually to destination" is the use of HDG on the MCP and Open Descent then this is the same thing as saying one had to use the steering wheel in one's car to get to one's destination and we all know how risky that is. If one selects a downstream waypoint to which one wishes to "go direct", one selects the waypoint, puts it into the requisite line and the airplane will turn towards the waypoint using the shortest route, left or right. In the Boeing, after selecting and placing the waypoint into the requisite line, one then must execute the selection, leaving the crew one more step to make sure that that is what they want. There is no "maybe" about this.

RR_NDB

I would add:

And assure the man-machine interface will always be able to HELP effectively the crew even when facing (all possible) extreme conditions. Allowing a FAST ("immediate", if possible) understanding of the problem(s) or threat(s).

In order, at least, to allow the very basic: Aviate and Navigate safely.

In a "graceful degradation" environment in order to increase chances ("giving" time) to succeed.

(*) The "effective aircraft" (System + crew) must always have (good) chances to "survive". Redundancy is the Key. "Sully" case may be is a good example on the need of a "pilot" up front. The System suffered a major failure. Compare "tiny ice crystals" with "flock of canadian geese"

Problem seems:

You need a System specialist and a Pilot. "Better" would be, first a Pilot and (last but not least) a System specialist.

The "microprocessor fired" the FE's. Is the automation (Super Systems) threatening the "pilot side" of the guys up front?

CONF iture

His words :
"The buffeting is a fairly low frequency vibration of the whole airplane, it shakes. It is described per the English term : it is deterrent, which means it is frightening. A sign that all pilots must identify to let them know : Do not go further.
it is impossible that they did not perceived it. You cannot miss it. It shakes badly."

ref this post
You most probably didn't mean ND for the AoA.
Isn't Stall Warning supposed to be lost with all ADRs selected off ... ?
In my experiment THS never did ... ?

Not any more.
At the initial aural stall warning, the airplane is still flying and considering as approaching the stall. The procedure is now clearly to lower the nose first and even to have to reduce the thrust in case of lack of pitch down authority.
The initial TOGA action has been banned. Minimizing altitude lost is not the priority anymore, reducing the AoA is.

PJ2

Quote:
Originally Posted by PJ2
Thirty seconds after the first Stall Warning passing through FL270 the AoA was 10degND
ref this post
You most probably didn't mean ND for the AoA.
Isn't Stall Warning supposed to be lost with all ADRs selected off ... ?

Wow..."10degND"! Shoulda caught it, thanks.
Quote:
Pitch slowly reduced to about 10degND still with full forward stick. As it was held the THS unwound and returned to normal settings.
In my experiment THS never did ... ?

Hm, I can't explain that without more information. Ours unwound normally and one could feel it in the response of the airplane.
Quote:
In the approach to the stall, the SOP is to apply TOGA power, lower the nose and minimize altitude loss. This still applies!
Not any more.
At the initial aural stall warning, the airplane is still flying and considering as approaching the stall. The procedure is now clearly to lower the nose first and even to have to reduce the thrust in case of lack of pitch down authority.

The initial TOGA action has been banned. Minimizing altitude lost is not the priority anymore, reducing the AoA is.

Yes, I do recall the change now - what I was trying to do was separate two phases here - the "approach to the stall", and the "full stall" obviously and had referred to publications before the change. Kind thanks once again for the correction.
Have you any thoughts on the altitude required to recover from the full stall? What was your experience in your own experiment?