Airbus crash/training flight

[SKS777FLYER]

In the 30 years I participated in recurrent training (ugghhh) at AMR, we did many many aproaches to stall in the simulators. Most usually, the instructor (sim instructor or CKA) would have the trainees level at perhaps 10,000 feet, handflying the sumulator. No auto-throttles. The sim instructor would set the power very low and tell the pilots to leave the throttles alone until the first indication of impending stall (ie, stick shaker) and not to trim the simulator. The recovery was to shove the throttles full forward/ ask for max power, keep the wings level, minimize altitude loss and fly out of the situation. The same thing was repeated in a left or right turn; trimming as necessary.

[Mad (Flt) Scientist]

Originally Posted by bearfoil

Smilin Ed

Just a quick question. You say the USN trains not to trim into a deliberate Stall. Did you mean inadvertently, or deliberately? I might sound pedantic here, but as a training issue, once all the other parameters are in place, including sufficient altitude to recover, why would an assist from the trim wheel be out of place? Especially across all types in the Inventory?

bear

If a type has been demonstrated to have acceptable stall recovery techniques when trimmed for initial speeds X through Y prior to slow down for a stall, but either not demonstrated OR demonstrated to have marginal or unacceptable characteristics when trimmed at speeds closer to the stall, it would be prudent to restrict training to the already demonstrated range of acceptable trim conditions.

Obviously the Part 25 regulations specify such a range, but it may be that on the specific type or types a similarly restricted range of trim conditions was applied.

[Smilin_Ed]

Deliberate Stalls

Just a quick question. You say the USN trains not to trim into a deliberate Stall. Did you mean inadvertently, or deliberately? I might sound pedantic here, but as a training issue, once all the other parameters are in place, including sufficient altitude to recover, why would an assist from the trim wheel be out of place? Especially across all types in the Inventory?

Others have already addressed this, however, when a pilot is learning about the stall characteristics of a particular airplane, he should trim for some speed higher than the stall speed in the configuration under examination. That way after the stall, or approach to stall, is experienced, the pilot needs only to let go of the controls and the plane will fly itself out of the stall. If you trim into the stall you might have a hard time getting the trim back to an acceptable level, particularly when the trim rate is slow. That's what happened here.

[Bis47]

Poor training philosophy

Hello

the pilot needs only to let go of the controls and the plane will fly itself out of the stall

Very bad training philosophy, imho.

If you teach stall recovery, teach it in realistic conditions.
In conditions less than ideal, in a context that includes some equipment malfunctions and/or pilots mistakes : those are the usual pre-conditions of a real accidental stall ...

If you demonstrate approach to stall recovery in "ideal" conditions, where is the interest? In ideal conditions, you won't approach stall ...

Instead, go as deep as possible into realistic teaching ... A simulator crash is not a drama ...

A light airplane is a good teaching tool too :
I can demonstrate trim upset in a C182 ... with different recovery options
And accidental vicious spin entry as well. I garantee that!

[bearfoil]

Bis47

I think that was the intent of my question. I know Smilin Ed wasn't referring to a baseline training syllabus when he suggested that trim avoidance was taught in all training.

Cross type is a bit of an oxymoron, especially when lack of training and misunderstood aerodynamic handling of a specific a/c has recently killed people.

As you say, training approaches differ on the level of experience and sophistication of the a/c in question. A small Cessna may need a full boot of rudder merely to get it to spin in the first place, while a lack of attention to HS trim can get TRE's and C/A's killed in a sophisticated transport. There is no universal training.

An A-4 needs Pitch input to launch. An F/A-18 doesn't allow the aviator to touch the stick at cat-stroke, and until positive Roc.

When the F-18 "departs" the pilot is required to release the controls until the "box" regains aero. An A-4 in departed flight requires pilot input. This accident is more about misunderstanding one's a/c and overconfidence than a/c unsuitability. IMO.

[jcjeant]

Hi,

Will not search and quote .. but in the FDR analysis .. seem's the pilot (s) regained full control on the plane for a little time (he was again flying normally) and again the plane departed from normal flight ...
So it's seem's the altitude is not the real problem (as the pilot(s) was able to recover) .. but something on the plane make it departed again .......
Just a tought.

[bearfoil]

"again the plane departed". If it has been discussed, I haven't seen a discussion related to g loading of the airframe and controls location or input/ location discrepant data. What does this a/c do when "protections" cross swords at inopportune times? What does she do when "g" is negative? The last I remember of the report has her leaving the cloud layer at 60degrees n/u. She entered the water n/d, what was the transition relative to Flight Law, "g", and AoA? Isn't that the fulcrum of the discussion? Did someone mention the lack of a published loading trace?

bear

[jcjeant]

Hi,

From the BEA final report (captures from PDF documents)

[Machinbird]

jcjeant

but something on the plane make it departed again

That something was full nose up trim with the aircraft in direct law.
That was the point of the A-4 Skyhawk recovery procedure 13 posts before this.
The aircraft was recovered to straight and level flight at 138 knots(from memory) but as it continued to accelerate the aircraft began to promptly pitch up, finally getting up to 57 degrees (also from memory). The pitch up was rather quick. The aircraft never regained auto trim although it switched to alternate law on gear retraction.

If the A-4 procedure is relevant to this event, which I suspect it is, the way to recover from this pitch up would have been to immediately reduce power and to slow the aircraft down promptly to prevent the nose from being driven higher. (Imagine doing this just after stalling your aircraft and recovering!) The intent of the procedure would be to gain time to understand what was happening. The trigger for such a procedure would be a pitch up related to aircraft acceleration.

I am fairly confident you do not have such a procedure in FCOM. And I'll wager that if we were to ask a group of Airbus drivers if they needed such a procedure, the answer would be negative.

[bearfoil]

Machinbird,
Many operating manuals direct increase of throttle w/o altitude gain through the shaker, prior to Stall (No Stall recovery is included, there is not supposed to be a Stall). If in direct Law, it would seem the same procedure 'after Stall recovery has been accomplished' might be indicated. With full n/u trim defeated with a return to Alt Law and gear retraction, my question is still, why did the FP "redepart"? Also, was there a "g" limit imposed on his unsuccessful pullup at the water's surface (Alt Law prot?). I've read the reports, and don't 'get' the sequence of extremely odd a/c behaviours. I also don't understand why the pilot should be necessarily frightened out of counter-intuitive recovery after Stall recovery; Flying is flight, at any speed. You are saying he demanded too much "wiggle room"?

perhaps I am just thick.

bear

[fdr]

LOC

Bearfoil

the initial pitch up is a direct response to the thrust application, and achieves a rapid pitchup rate being developed. During the pitchup, the pilot inputs full AND SSC elevator input, but the pitch couple of the thrust line defeats the available elevator authority in the full ANU stabiliser position which has been achieved due to the continued trimming during the deceleration which has occurred due to the underlying fault of the AOA probes being frozen from water ingestion through the seals. The aircraft (IIRC... haven't looked at the data for about 12 months...) exhibits lateral instability during this due to aerodynamic stall, coupled in part with the plots lateral control inputs (the system is not designed to be operating in this regime... and this area was the most interesting to assess if the aircraft fully complies with CS25 stability requirements, it being evident from a number of cases that the FBW aircraft can end up in aerodynamic stall contrary to the design specifications of envelope protection. A conventional control aircraft is required to achieve particular lateral stability and control requirements in a stall, whereas the certification of a FBW aircraft that "avoids" stall has alternative compliance criteria....)

The aircraft pitches up due to the out of trim condition and achieves a very low speed, high pitch attitude, but at a moderate AOA, until it finally runs out of energy and starts to fall off with both roll and pitch excursions, with initially increasing AOA. The ensuing flightpath increases CAS, and results in an increase in AOA again, leading to secondary stall... pitch attitude lowers and the aircraft enters a final dive.... with increasing CAS and increasing AOA again...

To regain control the crew needed to reduce the pitch up couple ie reduce thrust, or trim the stab forward from it's extreme ANU position (11u?... memory fading...), and lower the pitch attitude, possibly by rolling off. Ailerons are low effectiveness (and the data showed possible roll reversal IIRC... which is not too good for a CS25 certified aircraft... ), and judicious use of rudder sounds nice but is rather nasty in a swept wing jet transport (do not take the simulators behaviour in training as being representative of the control derivative in the stall regime... data from various events show that the aircraft can achieve very high roll rates ie 5-6 times higher than normal control authority when in a stalled condition). Judicious may be a great term in a subsequent inquiry but it is hardly precise in a high oscillatory loading, dynamic upset event. refer FAR25 CS25, and AC25.7A.

Power reduction at low altitude in a low speed condition is not going to be foremost in a pilots mind... and the rapidly deteriorating flightpath (which additionally result in g loads in the cockpit that are not simulated in any FFS) results in a high cognitive workload, and a likely massive spike in stress levels in the crew. The oversight of the trim state is not a surprise, and has been evidenced on numerous occasions on various types, (the A310 being over represented, as well as the A300-600...)

Humans are a bit like democracy, a lousy program, but far better than all the alternatives... The crew didn't go out planning to have a bad day, but the circumstances they were placed in technically coupled with their unfortunate ad-hoc attempt to achieve a desired "mission" goal placed them in a rapidly deteriorating situation where the time line, workload and stress resulted in their being unable to recover the SA that they had lost in the commencement of the procedure.

Any perception that aviation is a simple task fails to understand the facts of the physics of flight. The routine nature of RPT makes it easy to underestimate the underlying demands of operating high energy machinery in varying environmental conditions with varying levels of system serviceability.


Would think it appropriate for AI to add a CWS voice command such as: "MANUAL STABTRIM REQUIRED" or similar for degradation to direct law, instead of the fairly anemic ECAM messages which were patently not cognitively effective on the day.

RIP.


FDR

[BOAC]

Originally Posted by fdr

Ailerons are low effectiveness (and the data showed possible roll reversal IIRC... which is not too good for a CS25 certified aircraft.

- I cannot see how you can defeat basic aerodynamics? Roll reversal is a given, surely?

Rather than introduce another audio input into a by now highly stressed cockpit (what happens to the audio senses during stress??) why not adopt my suggestion of reminding the crew that they are trimming too far n/up (requiring an 'over-ride' action) BEFORE it all goes to pot? Quite a few of our recent pitch/power couple events might have been avoided.

[PBL]

Originally Posted by BOAC

another audio input .... (what happens to the audio senses during stress??)

Patterson on Auditory Warning Sounds, Phil. Trans. Roy. Soc B, April 1990

PBL

[bearfoil]

fdr

Thank you for a remarkably explicit and concise summary. I'm sure it's been covered, then, the occasional conflict between a complex system and aerodynamic basics. Without correct and basic information, situations arise rarely that challenge the pilot's training and experience. Masked over ports, hornets, birds nests, and moisture still plague the endeavour. I am somewhat curious about how the pitch/power couple and bunk trim did in this pilot.


bear

[fdr]

reversal certification reqts

Hi BOAC,

the immutable laws of physics exist however (roughly speaking) the certification reqts can be met by good design, or alternatively by a system that guarantees that the exposure to the condition does not occur....much....

As Richard P. Feynman wrote on the Challenger disaster:

"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled."


Feynman's Appendix to the Rogers Commission Report on the Space Shuttle Challenger Accident



The stall speed is defined as the minimum speed reached during the maneuver, except for those airplanes that require stall identification devices (see paragraph 29b(3)(iv), below).
(i) The pitch control reaches the aft stop and is held full aft for two seconds, or until the pitch attitude stops increasing, whichever occurs later. In the case of turning flight stalls, recovery may be initiated once the pitch control reaches the aft stop when accompanied by a rolling motion that is not immediately controllable (provided the rolling motion complies with § 25.203(c)).
(ii) An uncommanded, distinctive and easily recognizable nose down pitch that cannot be readily arrested. This nose down pitch may be accompanied by a rolling motion that is not immediately controllable, provided that the rolling motion complies with § 25.203(b) or (c), as appropriate.
(iii) The airplane demonstrates an unmistakable, inherent aerodynamic warning of a magnitude and severity that is a strong and effective deterrent to further speed reduction. This deterrent level of aerodynamic warning (i.e., buffet) must be of a much greater magnitude than the initial buffet ordinarily associated with stall warning. An example is a large transport airplane that exhibits “deterrent buffet” with flaps up and is characterized by an intensity that inhibits reading cockpit instruments and would require a strong determined effort by the pilot to increase the angle of attack any further.
(iv) The activation point of a stall identification device that is a strong and effective deterrent to further speed reduction. Paragraph 228 of this AC presents guidance material for demonstrating compliance of stall identification systems with the regulatory requirements of Part 25 of the FAR.


d. Stall Characteristics.
(1) Background. Since operational pilots may not be required, or trained, to fly to an angle of attack beyond that for stall warning, any exposure to the behavior of the airplane in an actual stall would be both unexpected and unfamiliar. Therefore, to assure a safe and expeditious recovery from an unintentional stall, it should not require any unusual piloting technique to successfully demonstrate compliance with § 25.203, nor should it require exceptional skill or repeated practice by the test pilot. The behavior of the airplane during the stall and recovery must be easily controllable using normally expected pilot reactions.

Section 25.203(b) states that “the roll occurring between the stall and the completion of the recovery may not exceed approximately 20 degrees” for level wing stalls. In level wing stalls the bank angle may exceed 20 degrees occasionally, provided that lateral control is effective during recovery.

The definition of adequate controllability is specifically described in the context of abnormal aerodynamic configurations:

“Adequate controllability” means that it is possible to produce and to correct pitch, roll, and yaw by unreversed use of the flight controls, and that there are no uncommanded airplane motions due to aerodynamic flow breakdown. If stall warning is used as the end point, then it should be demonstrated that the airplane is safely controllable and maneuverable when flown at the recommended operating speed.

Curiously, there is no testing reqt for the out of trim case to look at the condition that occurs in a stall recovery with trim/power set in an adverse configuration, nor is there an A-PC HQ evaluation required in this case. It would appear that the potential to place the A320/A310/A30-600 out of trim with large thrust couples would be worthy of evaluation, given the adverse outcomes that have occurred (B737 on occasion as well...).

FDR

[BOAC]

Despite all the fine quotes, fdr, I will doubt you (or anyone) will ever manage to change the roll properties of a fully stalled conventional wing - as someone once said "for nature cannot be fooled"

[Mad (Flt) Scientist]

@BAOC - true, the key then is to attempt to prevent that fully separated wing stall from occurring. Getting part of the wing to stall, while keeping other bits flying, in a fashion that both warns (deters) further AoA increase while retaining the control to recover the aircraft is the basic goal of any naturally stalling wing design.

Since some people do achieve such naturally stalling wings for certification, and others manage to place their stall identification or deterrence devices (e.g. stick pushers) at higher AoAs than the onset of the aerodynamic stall, clearly it can be done.

[MountainBear]

the key then is to attempt to prevent that fully separated wing stall from occurring

The problem is that there is a basic and unavoidable trade-off between prevention and cure. The human body is designed to avoid falling, but once you go off that cliff human's don't sprout wings. Likewise, a bird is designed quite well to deal with falling but doesn't have many features to prevent it from falling in the first place.

The problem then is that as you increase the complexity of the systems to prevent that wing from stalling you also increase the risk that when that prevention fails (and it will inevitably fail sooner or later) there is no cure left.

Mother nature cannot be fooled and one thing it's foolish to believe is that you can design a perfect system. It's always a question of trade-offs, of robbing Peter to pay Paul.

[p51guy]

I agree with fdr. If you get into an aerodynamic wing stall in the lower altitudes recovery should be possible by first not overusing thrust and using elevator and stab trim to get the nose down again. I believe it would work in all the Boeings I have flown. Never flew the bus.

[jcjeant]

Hi,

I believe it would work in all the Boeings I have flown. Never flew the bus.

Seem's for this particular Airbus it unfortunately not worked