Originally Posted by
qquantum
SLFinAZ – you sound like the NTSB.
As expected, the BEA Report is comprised of carefully crafted smoke — placing blame on the pilots of course — and distraction from the real issues of airplane design and regulator oversight failure that truly were the causes of the accident:
A lot of stuff from one so new here, maybe you haven't read all the previous threads on the issues you raise, and you clearly haven't read the report.
In all, a "politically correct" report that omits analysis of significant factual information,
The facts are there, you just seem to have omitted to assimilate them.
1. The loss of airspeed information caused a change in the FCS mode, which the pilots seemed not to notice, as they made no mention of it;
False. Mode change noted and read out. Read the CVR in the report.
2. The FCS Mode change resulted in loss of the envelope speed and angle of attack (AOA) limits — no stall protection — no overspeed protection — no bank angle/pitch angle protection — again the pilots seemed not to notice, as they made no mention of it;
False. As above.
3. The inexperienced FO, who was flying, made a radical pitch control input (full aft stick), which was not noticed by the more experienced FO, as he could not see the opposite control stick in the darkened cockpit
False and false. Full back stick came much later when stalled. Initial climb was noted by PNF and he instructed corrections.
4. This drastic control input caused the trimmable horizontal stabilizer (THS) to automatically apply full nose‐up trim
False. The THS didn't wind up until they were already stalled, and there is no evidence autotrim failed - with nose down control inputs the THS would have followed allowing recovery (plenty of SIM sessions and theoretical models covering this on the tech log thread).
5. Examination of the Flight Data Recorder (FDR) data, (Final Report, Appendix 3) indicates that, incredibly, the stall warning ceased (as designed) while the airplane was still in a deeply stalled condition, then reappeared when recovery action was attempted, adding even more confusion to the situation.
Irrelevant - there is no evidence that stall was ever diagnosed - which is a pre-requisite for a recovery attempt. There was no recovery (from stall) attempt.
Conditions of high altitude stall and reversion to Alternate Mode have been very rarely encountered in airline operations, leading to a degree of complacency
Stall - rare. UAS and alt-law - not rare, known and discussed between mfr, airline and regulators. Chronology is in the report.
Two possibilities:
1. Crew was abnormally (in)competent - and that's the cause.
2. An averagely competent crew will turn uas + alt-law in cruise into a stall with an unacceptably high probability (say > 1 in 100). Cause - who assessed that risk (if anyone) and did they get it right ?
Design Flaws
1. Failure of the FCS to remain in the Normal mode (with envelope protection) resulting from the loss of a single parameter (pitot‐static airspeed).
Designed as required for certification. AP & protections shall not operate on known-bad data.
If the design had incorporated synthesized speed – from GPS, IRS – which would be accurate enough – the pilot would not have been placed into
Airspeed != ground speed or GPS speed or inertial.
needless to say, the presentation of angle of attack (AOA) information (now planned for inclusion on Airbus airplanes) would have eased the burden on the pilot.
AOA already available on PFD via flight path display. Separate AOA already an option. Airlines do not take it, allegedly to avoid expense of training pilots to use it.
2. The design of the primary flight control controller such that it is possible for a pilot to make a sustained incorrect or dangerous input, without the awareness of the other crewmember(s).
Not proven - PNF was aware, see report.
Assessment of whether input is incorrect or dangerous is dependent on diagnosis / awareness of situation, which was not present. Pull back into stall happens on other control systems (see Colgan, Ethiopian, Birgenair etc.). Common thread is not control type but the lack of diagnosis of stall.
It can be readily appreciated that an FCS control design in which both controllers move in unison at all times (as in Boeing, MacDonnell Douglas, and most other airplanes) would have enabled instant detection (tactile) by the senior pilot (along with the probability of a sharp rebuke) and subsequent recovery of the airplane.
Not without diagnosis of stall, hearing and understanding the warning etc.
There are more Boeings at the bottom of the ocean due to this issue than 'buses.
now, however, the required technology has been available for several years, and because tactile communication between pilots is of such vital importance, parallel control should become mandatory on all future sidestick designs.
AB sidestick design pre-dates the technology. Future sidestick designs from other mfrs might go a different route. AB sidestick very unlikely to change as it is proven and changing it would break the cockpit commonality across the range.
3. The ability of the FCS to allow the Trimmable Horizontal Stabilizer (THS) to move to a configuration that renders the airplane essentially unrecoverable, without appropriate warning to the pilot.
Trim wheel is visible and moves. Should it have additional audio warning - maybe, but consider they didn't hear (or process) prolonged stall warning.
Further, when AOA becomes greater than 30°, or Speed is less than 60kt, auto THS trim is turned off.
Reference ? Report contains no indication of autotrim ceasing to work.
4. The FCS design which allows the primary stall‐warning sensor to be declared invalid (when it is still capable of providing a correct indication) based upon another parameter (speed), of questionable validity.
Wrong. FCS has no bearing on this. ADIRU does. Be aware that other types also have ADIRUs...
SW computer ceased warning because its input AOA went "invalid" - airspeed did not factor (except at the ADIRU).
When the senior co‐pilot took control (although the junior FO did not relinquish it) and pushed forward on the stick in an attempt to recover, the stall warning again sounded – leading him to believe that his control input was not correct, thus causing him to release pressure on the stick to get rid of the stall warning.
No indication that stall was ever diagnosed.
By the time this warning issue occurred the crew had already ignored 1min of stall warning and had got to AOA > 40deg. They were far beyond where any sane test pilot would go in conditions no test pilot would test in.
Had they started looking at an analogue AOA gauge at that point it would have been pinned at max and appeared "stuck" anyway.
It's an interesting design issue, it's unlikely to be specific to the 'bus, but it's IMO academic for 447, by that time the outcome was already decided.