Wednesday, February 16, 2011

AF 447 Air France flight lost over the Atlantic

I just learned that NOVA will have a special tonight on the AF 447 flight. There are so many interesting aspects to this puzzle. Primary amongst them is the Airbus concept of a computer flying the aircraft and over-riding pilot input in certain modes. Also involved are sensor design issues, conflicting data resolution issues, pilot attention issues, international cooperation issues, search operations issues and use of maintenance data as an alternative to the unrecovered Black Boxes.

I'm uncertain if the plane deviated for severe weather or not. I think the primary problem was overspeed resulting in a deep stall at high altitude resulting in a near vertical descent with little forward speed. It would be a problem of how to evaluate fluctuating data from multiple sensors.

The flight into a severe thunderstorm may well have taken place although the pilots may have believed they were above the storm. In most stalls the rule is simple: lower the nose immediately and add power immediately. This simple rule usually results in greater air flow over the wings to generate lift. Wings are usually designed to stall from the inboard area first so that the pilot has time to react before the stall reaches the outboard areas of the wing and the control surfaces become stalled. In a deep stall, moving the control surfaces has no effect because there is no airflow over the control surfaces. Adding power may be a problem if there was a sudden and unexpected ingestion of super-cooled rainwater into the engines. If it was not just pitot tube icing but an engine flame-out, no power may have been available to the pilots even if in their information overload state of mind they realized they desperately needed to add power.

Analysis of the flight from the ACARS maintenance data transmissions is not ideal. The information is designed to give maintenance crews an indication of what systems will need to be tested when the plane lands and what tools and parts should be on hand for the task. Such maintenance data is not an alternative to a Flight Data Recorder. There are also problems since the ACARS messages are held in a one minute suspense window and are grouped by item and priority. Certain message evaluation rules are clearly compromises. A fault detected by a system is considered more serious if it relates to a separate system than if it is a self-detected internal fault. A fault is deemed to be hard if it persists for thirty seconds whereas a fault is deemed soft if it is detected for less than thirty seconds. Such rules do make sense in most circumstances.

The loss of satellite uplink carrier signal during a critical time is one argument against replacing physical black boxes with constant electronic transmission of the FDR and CVR data. The cost of such transmissions is a major problem also.

I'm sure it will be an interesting program and I look forward to viewing it. I hope all my many readers (LOL) will enjoy it.

By the way, the most relevant incident to compare AF447 with is the Gimli Glider incident. That plane ran out of fuel half way through its flight and all the design errors, pilot errors and ATC errors were countered by one thing alone, pilot expertise. The pilot had a hobby of flying gliders, so when he suddenly found himself flying a huge ungainly glider full of passengers, he still knew what to do and how to do it. He had flown gliders before. He knew to make his S turns toward his objective and not lose sight of it. He knew how to do skids. He was flying the airplane rather than monitoring computers but everything turned out okay because he was used to actually flying the airplane rather than monitoring computers. That is a rare skill nowadays.

With AF447 when things turned so bad that the computer entered Alternate Law mode the computer would no longer override disapproved pilot inputs and would allow the pilots to fly the aircraft as they saw fit. Unfortunately, the pilots were not very experienced at actually flying the aircraft and were not able to do it in situations of high informational overload, high insecurity as to the reliability of their instruments and unusual attitudes. When Fly-By-Wire ended and Alternate Law started, the situation was already critical and the pilots desperately needed a learning curve. Unfortunately, there is no learning curve during such an event. If you've entered even the fringes of a supercell at night and are experiencing electrical overloads, have circuit breakers popping all over the place, are losing some instruments and have utterly lost faith in all your instruments, the situation is hopeless for a pilot who really doesn't know how to fly an airplane without a computer.

Every flight using a computer to change altitude saves on fuel costs over a heavy handed pilot trying to manipulate the throttles, but every such flight deprives the pilot of real flying experience. Just as the Gimli Glider incident involved a sudden transition the AF447 undoubtedly involved a sudden transition even though we don't know the cause. Initial lightning seems insufficient. Iced-up pitot sensors can have any number of causes. Auto Pilot and Auto Thrust are different systems and some confusion may have existed as to speed selection and the rapidity of speed adjustment attempts that were required.

If there was a deep stall, I just wonder what attitude is required for it. Note: A Deep Stall is one wherein no manipulation of power or control surfaces will have sufficient effect. A pilot in the deep stalled plane is as utterly helpless as a pilot on the ground who might be observing the incident. The pilot on the ground will push a non existent throttle lever, the pilot in the plane will push an actual throttle lever however the effect will be the same. A pilot on the ground will push an imaginary steering yoke, a pilot in the plane will push a real steering yoke yet the results will be the same. No effect whatsoever. A deep stall with or without a flat spin would mean a high speed vertical descent with little forward momentum.

NOTE: On Feb 28 2011, the Alucia arrived, fresh from a Seattle overhaul, at the Pacific waiting area for transiting the Panama Canal to begin its three phase search for the Black Boxes using remote controlled submersibles.

NOTE: This Phase 4 search continues. IF there is a Phase 5, Recovery Phase, a different ship and different equipment will be used. So far, no debris field has been discovered much less the "black boxes". (April 2nd, 2011).

Debris Field found. Learmount of Flight International being quoted on theory of crew fatigue induced fixation on minor problem while plane was in an unnoticed gentle descent resulting in a collision of an intact and controllable plane with the sea. He does not believe there was any deep stall. Merely pilots having lost so many instruments that they did not believe what few instruments remained.

9 comments:

FleaStiff said...

France just indicted Airbus for manslaughter. The preliminary filing by the Magistrate is ill-timed since the research vessel Alucia is about to conduct a search for the FDR and CVR.

FleaStiff said...

It is difficult to know just what acts support the manslaughter charge. Some allege an onboard weather radar would have detected a small cell in their path but masked the larger cells behind it.

Its hard to imagine that a deep stall resulted from contact with a supercell. The plane impacted the ocean with control surfaces still attached. Useless perhaps, but still attached.

The crux of the problem may be that Alternate Law is simply too novel a situation for the pilots but one wonders if that is something Airbus should be blamed for.

FleaStiff said...

A series of ignored complaints about the pitot heaters and a series of improper servicing of the pitot heaters may indeed be the basis of the proposed manslaughter charge, but I wonder if anyone ever considered the difference between loss of the pitot heater data and conflict over the pitot heater data?
It was a fly by wire system that appears not to have dealt with that conflicting data properly.

FleaStiff said...

Alucia is on station and the 16 week project is underway.

Very little notice is being given to the press for briefings and most of the press coverage has been little more than stock footage of the research vessel and the remotely controlled search submarines.

The Brazilian press is not particularly noted for precision in its reporting but it is hoped that there will be a focus on technical updates as the search progresses.

FleaStiff said...

Initial area of Resumed Search is at the 20 NM mark NW of the LKP. If throttles were retarded transition from Mach.82 to Mach.53 would take two and a half minutes at which time minimal buffeting would be encountered. This would have allowed travel to just two miles beyond the search area of April 2nd 2010. Search will start there and work its way southwards as necessary.
German press covering this more precisely than Portugese press, but Portugese press appears to be doing well overall.

FleaStiff said...

There is a great deal of talk about the coffin corner (the difference between the minimum speed and the maximum speed) and its various precise limits. The problem with looking at this as a simple stall is that the wreckage and the maintenance data that was broadcast indicate a deep stall took place wherein there was little if any forward motion and the major vector was downward. Instead of explosive decompression at high altitude (which is more annoying than anything) I think that plane hit relatively intact with all control surfaces attached but ineffective.
Being thrown into Alternate Law mode with no warning may have provided a cognitive overload for the pilots but there should have been enough time for the pilots to do something and do it right. I can not believe that they were ignorant of what to do so the only conclusion is that they were unable to change the attitude of the aircraft and that means all control surfaces were fully stalled.

Anonymous said...

dont you mean underspeed resulted in a high altitude stall

FleaStiff said...

As mild buffeting started and perhaps instruments were still being relied upon an underspeed situation would have resulted in an immediate attempt to add power. With low-slung engines and a aft Center of Pressure there would have been a pitch-up despite pilot commands. If the plane then found a severe downdraft on the outer edge of a storm cell there could well have been a rapid descent in a stalled attitude.

Nalliah said...

This Airbus A330-203 did not have multiple independent systems for detecting speed of the aircraft such as a GPS based system that would at least cross check the readings being given by the pitot tubes and then provide a cockpit warning that the airspeed could be wrong, or another safety mechanism whereby the pitot tubes are heated as long as this would not impact the reading so that ice could not occlude them.

The accident was caused by the co-pilot induced Deep Stall condition and remained in that condition until impact.

To recover from deep stall is to set engine to idle to reduce nose up side effect and try full nose down input. If no success roll the aircraft to above 60° bank angle and rudder input to lower the nose in a steep engaged turn.

Pilots lack of familiarity and training along with system malfunction contributed to this terrible accident. Also the following contributed to the accident

(1)the absence of proper immediate actions to correct the Deep Stall

(2) Insufficient and inappropriate situation awareness disabling the co-pilots and the captain to become aware of what was happening regarding the performance and behaviour of the aircraft

(3)lack of effective communication between the co-pilots and the captain which limited the decision making processes, the ability to choose appropriate alternatives and establish priorities in the actions to counter the Deep Stall

During most of its long descent into the Atlantic Ocean, Airbus A330-203 was in a stalled glide. Far from a deep stall, this seems to have been a conventional stall in which the Airbus A330-203 displayed exemplary behavior. The aircraft responded to roll inputs, maintained the commanded pitch attitude, and neither departed nor spun. The only thing the Airbus A330-203 failed to do well was to make clear to its cockpit crew what was going on.Its pitch attitude was about 15 degrees nose up and its flight path was around 25 degrees downward, giving an angle of attack of 35 degrees or more. Its vertical speed was about 100 knots, and its true airspeed was about 250 knots. It remained in this unusual attitude not because it could not recover, but because the co-pilots did not comprehend, in darkness and turbulence and amid a tumult of conflicting warnings of mysterious system failures, the actual attitude of the aircraft. The co-pilots held the nose up. If the co-pilots had pushed the stick forward, held it there, and manually retrimmed the stabilizer, the airplane would have recovered from the stall and flown normally.

Practicing recovery from "Loss of Control" situations and improve flight crew training for high altitude stalls (simulator training usually has low altitude stalls which are significantly different due to energy status of the aircraft) should become the mandatory part of recurrent training.