Airbus crash/training flight

[FlightlessParrot]

A question, if I may.

Assuming they were checking alpha-floor protection, what would have been expected to have happened, if all had gone well?

[Massey1Bravo]

By looking at the graphs, it looks like the Airbus got itself into direct law and with full thrust, started pitching up uncontrollably even as the captain pushed his sidestick fully forward.

After the first stall the stabilizer position is 12 degrees up while the elevator position is 12.5 degrees down. Is this normal?

[RevMan2]

Transfer this incident to a business environment.

You're a software provider and you have a contractual obligation to deliver a core business application to your customer at a specific time in a sunset scenario. (Old application HAS to be switched off then and then, new application HAS to be seamlessly introduced)

Together with the customer, you define the business (use) cases to be validated, develop a test plan with adequate time buffers and perform and document the tests.

At some point during the test plan, you determine that you're unable to complete all acceptance tests without jeopardising your delivery date. The reasons are irrelevant - external factors, poorly structured test plan, repeated testing in order to validate bug correction.

The fact is that you're left with some minor uncritical business cases to test, but in testing them, you're running the risk of not being able to compile the application in time for on-line release.

What do you do?

Press on regardless and risk a catastrophic commercial outcome if the final compilation doesn't work first time (with no second chance), or release the application in a stable state and continue testing off-line?

It comes down to risk assessment - is the risk we're engaging in commensurate with the benefits?

In this case: was the acceptance test required to be concluded with the touch and go at Perpignan or could the low speed tests have been performed enroute to EDDF?

[NigelOnDraft]

DC-ATE

However, if there had not been a need to "check the alpha floor protection" in the XL aircraft, there would not have been an accident.

If you actually read the report you will see there was no need to check the Alpha Floor

The Alpha Floor check is an Airbus Test Pilot procedure on pre-delivery flights only. It was to be conducted in Normal Law only, and above FL100. In Altn/Direct Law the test was meaningless...

For whatever reason, this one was conducted by a non-Test pilot, at 3000', and in a "normal aircraft" i.e. Direct Law.

Did you, in your DC-8 time, deliberately stall the aircraft, with Gear and Flaps fully deployed, and hold it into/through the stall with full aft stick, at 3000'? In a swept wing aircraft the results will be interesting

I need to re-read again, since I am not sure they intended to be in Altn/Direct Law for the stall (although the signs are there, least of all the Stall Warning). The AoA probes seem to possibly have some bearing on the sequence of events (but maybe not the outcome).

However, I might hazard 2 quick conclusions:

1. FBW is somewhat credited with at least avoiding this aerodynamic manoeuvre - when it works
2. The report does quickly home in on the the root cause of the accident - manoeuvring outside a defined requirement / limitations, and the potential (unanticipated) consequences of doing so
3. There is little "urgency" for the full report (despite the baying on here) since the accident appears to have little applicability to Line Ops

NoD

[Safety Concerns]

There is no urgency here as the design is sound and as NOD says there will be little to no impact on operations.

We are looking at human error.

I wish the dinosaurs would go away, if this was printed on paper how would we justify erasing a rain forest printing some of this drivel.

Even DC8's stall. Incredible but true DC-ATE

[Clandestino]

The alpha floor test was not done in direct but in normal law - you can see it clearly in the graphs. There was check of direct law earlier in the flight, but it is clearly stated that both FACs were turned on afterwards. However, as the aeroplane slowed down, AoA values were stuck at unrealistically low values - I suppose that BEA is focusing its investigation on what caused them to stick. Seemingly the FACs faced with conflicting AoA and air data shut themselves off just as the aeroplane was about to stall, turning control law into direct, with pilots having: full control over flying surfaces, no restrictions and no protections.

After the first stall the stabilizer position is 12 degrees up while the elevator position is 12.5 degrees down. Is this normal?

In direct law the only control over "trimmable horizontal stabilizer" is via pitch trim wheel. I'm a bit puzzled by THS staying stuck after pitch law was back to alternate, but I'll wait for further reports to enlighten me.

Also, it's really hard to believe that people are flying around in airplanes designed so that the pilot(s) can't instantly get control from a bunch of computers and fly the thing out of danger.

Of course it is hard to believe an untruth. The confused computers handed over full and complete control to surprised crew at just the wrong moment and crew's failure to exercise it may be eventually attributed to insufficient time (=altitude) to recover.

Sometimes the truth hurts, I know.

Only if it should.

[Christodoulidesd]

Instead of blaming the FBW systems or turn this into yet another (sigh) AvsB war, understand that there was NO NEED for such a test and certainly not at such a low alitutude!

Quoting from avherald.com

"The BEA reported, that that low speed test is expected to be performed at FL140. The computed speeds for full configuration would have been a VLS of 123 knots and a minimum speed of 107 knots."

So quit being close-minded and think first.
Cheers,
D

[bsieker]

Pitch trim

As I've read the report so far.

- FACs 1 and 2 FAIL status was recorded only between 4 and 8 seconds after the stall warning was recorded.

This is very strange if the FACs really failed late, because with all flight control computers working (no other shows FAIL on the FDR), the aircraft should be in normal law, and there should be no stall warning.

IF FAC failure is recorded with some delay, the following becomes plausible:

- Aircraft at high angle of attack (well into the ALPHA-PROT regime, but still short of ALPHA Floor)
- low airspeed
- No stall warning, as aircraft in fully protected mode ("normal law")
- both FACs fail (within a single FDR data sampling period)
- Aircraft now in direct law (gear down), with no protections
- Stall warning sounds, as flight parameters are already "close to stall", and protections are lost


At least two factors making recovery very hard or impossible:

- Pitch trim at full nose-up. Airbus FBW aircraft autotrim pitch loads in normal law, even without autopilot.
- Low altitude.


Given the previous comments about covering air data orifices and vanes prior to painting/maintenance, one can speculate that something went wrong there, other than that it seems very unlikely that both FACs would fail at the same time.


Bernd

[NARVAL]

"For you guys that do not understand Airbus FBW, DIRECT LAW flies just like a regular airplane, no protections."
Guiones, forgive me, but you lose autotrim, which makes quite a difference!
If you use full thrust then idle power etc...the plane cannont be controlled by the sidestick (elevator lacks authority if you do not manually use the trim wheel...and on the A320 that is not a reflex action)
No regular airplane there if you have not said during briefing prior to those manoeuvers "we shall lose autotrim, I will use the trim wheel to control attitude...)

[framer]

RevMan

Transfer this incident to a business environment.

No thanks, it may be gratifying for you to liken it to your particular area of expertise but it is not really relavent.
Although some of the human factors in your business environment scenario are present when flying an A320 with four people on the flight deck from two different airlines, and two different countries, communicating with someone from a third country through VHF radio while moving at 5km per minute.......there are other human factors that are not. It is not relavent because yours is a much simpler scenario with much less serious ramifications resulting from poor decision making.

It comes down to risk assessment - is the risk we're engaging in commensurate with the benefits?

Do you think perhaps that with 30,000 odd hours of flight experience on the flight deck that day, that the concept of risk management wasn't completely new to them? Maybe management should come down to the simulator and brief airline captains on risk assesment when they are doing their recurrent checks. Sounds like they got it all figured out.

[Flare-Idle]

Low-Speed Tests, Altitude & Training...

There is a need for standards and training requirements from the authorities for operators/crews performing acceptance flights. We used to discuss/design our acceptance flight profiles with the help of the aircraft manufacturer and its test pilots. Our small team of acceptance flight crews underwent several special trainings and recurrent in the SIM with the acceptance program to be flown real later on. Excellent technical knowledge of all aircraft systems was required. Of utmost importance was the close cooperation of the acceptance pilots with base maintenance during the ground time of the aircraft and prior the test flight for any special items to look for during the test program. We had an internal standard and qualification program in order to "certify" new acceptance pilots.
On our non FBW aircraft, acceptance flights were allowed day only and special WX minima were defined. Low-speed tests were always performed in an altitude box between FL150 and FL190 in close coordination with ATC.

[DC-ATE]

NigelOnDraft -

If you actually read the report you will see there was no need to check the Alpha Floor

I did read the report. I merely stated that had this not been a FBW aircraft, there would not have been this check in the first place. The fact that they might have done the check at the wrong altitude is another issue.

I have not fully stalled a DC-8 but did a 737 (-200). You are right; it's
not very comforting. I've posted my experience with that situation. I flew the DC-8 after the 737 and the 737s were the last airplanes that had actual training in flight. But, I'm sure others stalled the -8 at the time. The whole idea is the "recognition" of a stall and to react accordingly, not to experience a stall.

Like I've said in the past; I know nothing about FBW except what I read here and elsewhere. I don't know one "Law" from the next. It just makes no sense to me. And it is NOT a Boeing vs Airbus thing either, as Boeing is into FBW now as well.
================

Safety Concerns -

Even DC8's stall. Incredible but true DC-ATE

My what profound statement. ALL airplanes are capable of stalling. Well, actually, the 1049 Connie would just kinda sit there shaking trying to stall, but had a hard time of it. The whole idea is not to get in that situation.

We are looking at human error.

Very possible. I've said that about the majority of aircraft accidents even
being a pilot myself. The problem is accidents where the pilot is unable to
defend himself or relate what really happened.
==================================

Christodoulidesd -

Instead of blaming the FBW systems or turn this into yet another (sigh) AvsB war.....

As I have repeatedly stated; I am NOT trying to pit Airbus against Boeing. That would be kinda hard to do, right; being as how Boeing is into FBW now as well. All I've stated is that WITHOUT FBW, perhaps this accident would not have occured because there would not have been that check. You cannot argue that; whether the crew was doing this check at the wrong altitude or not.

[Engine3firehandle]

Airbus Flightcontrol laws

Just to explain it in a few words so that most of the readers and not airbus pilots understand.

1. Normal law: You got all the protections, highspeed, slowspeed protection, etc..

2. Normal law, protections lost:
You only got the markings on the PFD, but there will be no intervention from any flightcomputer if you go over a limit in bank, pitch or speed. ( hope this is right, a few years back that I flew it )

3. Direct law. Now the plan flys direct inputs from the stick and any smoothing is gone. You can do a roll, or whatever...

Please, this is a short answer in three sentences to the not airbus flyers and of course there a little things that have to be noted too, but are going very much into the detail. But I used to explain it in the beginning to my students like this, so they had a first idea about the system.

[guiones]

"Guiones, forgive me, but you lose autotrim, which makes quite a difference!
If you use full thrust then idle power etc...the plane cannont be controlled by the sidestick (elevator lacks authority if you do not manually use the trim wheel...and on the A320 that is not a reflex action)
No regular airplane there if you have not said during briefing prior to those manoeuvers "we shall lose autotrim, I will use the trim wheel to control attitude...)"

Narval, what the heck do you fly; anyone that flies Airbus FBW knows that during DIRECT LAW you will loose AUTO TRIM and there is a big anouncement in amber right in front of you to remaind you. Regular airplanes do not have auto trim !!!

The main point you are missing is the test was for protections related to NORMAL LAW, as soon as the A/C went to AlTERNATE LAW the test should have been discontinued; without mentioning again the altitude at which it was perfomed.

G

[kwh]

SLF writes...

OK, I'm not a pilot, but it seems to me that if you are testing something, anything, on an aircraft or on anything else, it has to be a possibility that the thing you are testing will fail to perform as expected. Otherwise, what is the point of carrying out the test in the first place?

It would seem to my simplistic mind that testing the brakes on my freshly repaired car is best not done while approaching a brick wall at 80mph, and that testing the stall protection on an aeroplane is best not done very close to the ground.

Is that an overly simplistic viewpoint?

[NigelOnDraft]

DC-ATE

I did read the report. I merely stated that had this not been a FBW aircraft, there would not have been this check in the first place.

I suspect there would... it just would have been the Stick Shaker or whatever Stall Warning system(s) were fitted. That is all they were "checking" - or rather, that was the intention of the check which they seemed to wish to perform, but under inappropriate/incorrect circumstances and requirements...

NoD

[NigelOnDraft]

Those are actually examining the report, I've taken a closer look... as others did above.

1. The approach to the stall appears to be in Normal Law.
2. Stall Warning in fairly steady state flight. This should have been a "heads up" things were going wrong... you should not get a Stall Warning in Normal Law [15:45:05]
3. Reaction appears to be TOGA ~2s later [15:45:06/7] followed by Fwd Stick [15:45:09]
4. FACs fail a couple of seconds later (or were switched off)... [~15:45:10]
5. A couple more seconds later aircraft goes to Direct Law (as one would expect with 2 FACs failed and Gear Down) [15:45:15]
6. Up to 15:45:21 the Pitch seems OK (decreasing slightly all the time since the stall warning), but from then on aggressively rises to ~60nu. Ful Power was achieved ~ 15:45:12]
7. Some interesting Wing Rocks in the above timings.

But what happened from 15:45:21? At this point speed had increased, wings were approx level, pitch was reasonable, full power had been on for some time, full fwd stick (but full aft stabiliser), ailerons were neutral...

I will await the BEA conclusions / proper interpretation of the above with interest... Of course, despite the "aerobatics" it had recovered from the stall etc. before impact, but height was insufficent to recover from the dive

NoD

[DC-ATE]

NigelOnDraft -

I suspect there would... it just would have been the Stick Shaker or whatever Stall Warning system(s) were fitted. That is all they were "checking" - or rather, that was the intention of the check which they seemed to wish to perform, but under inappropriate/incorrect circumstances and requirements...

Quite possible but of course, they would not have (or should not have!) benn testing a stick shaker at that altitude. Perhaps they were testing something else. We'll probably never know, because once again, the crew is not around to testify. Sad.

[NigelOnDraft]

Quite possible but of course, they would not have (or should not have!) benn testing a stick shaker at that altitude

Ditto Alpha Max... We pretty much know from the report that is what they were doing and that the only schedule requiring this was the one for Airbus Test Pilots @ FL100+... HASELL checks anyone

NoD

[night_flight99]

DC-ATE,

So what you are saying is that there is an appropriate part of the envelope to conduct these sort of tests. Thats surely true even of fbw aircraft. The design of modern transport aircraft has moved on. Lots would argue that this has been achieved successfully and that the modern product is improved.

Having flown products of Boeing, Fokker, MD and Airbus I would strongly argue that is the case!

Even in its most degraded state this fbw airbus has at least the protections afforded by conventional aircraft. Regrettably it would appear from the initial report that a test was being undertaken at an inappropriate altitude (and as you rightly point out you wouldn't conduct stall tests in a conventional swept wing jet at that same part of the envelope).