Air France pilots battled for 15 minutes to save doomed flight AF 447

Telegraph ^

[traumer]

Air France pilots battled for up to 15 minutes to save the doomed flight that went missing over the Atlantic this week, electronic messages emitted by the aircraft have revealed.

Details have emerged of the moments leading up to the disappearance of flight AF 447 with 228 people on-board, with error messages reportedly suggesting the plane was flying too slowly and that two key computers malfunctioned.

Flight data messages provided by an Air France source show the precise chronology of events of flight AF 447 before it plummeted into the sea 400 miles off Brazil on Monday.

These indicate that the pilot reported hitting tropical turbulence at 3am (BST), shortly before reaching Senegalese airspace. It said the plane had passed through tall, dense cumulonimbus thunderclouds.

At this stage, according to a source close to the investigation cited by Le Monde, the Airbus A330-200's speed was "erroneous" - either too fast or too slow. Each plane has an optimal speed when passing through difficult weather conditions, which for unknown reasons, had not been reached by flight AF 447.

Airbus is expected to issue recommendations today to all operators of the A330 model to maintain appropriate thrust levels to steady the plane's flight path in storms.

At 3.10am, the messages show the pilot was presented with a series of major failures over a four-minute period before catastrophe struck, according to automatic data signals cited by the Sao Paulo newspaper, le Jornal da Tarde.

At this time, the automatic pilot was disconnected – either by the pilot or by the plane's inbuilt security system, which flips to manual after detecting a serious error.

[expatpat]

....the weather been heavy enough to attenuate it Yes, see my post #112. However, that effect is well-known and the cell would still be visible and should have been avoided.

Re 153, I did miss your point.

[expatpat]

The ‘visible horizon’ is much further away if the observer and the object are both at 35,000ft, of course. If my calculation is correct, I get about 700 miles for the distance the two could see each other if the air were clear enough.

[NittanyLion]

The ‘visible horizon’ is much further away if the observer and the object are both at 35,000ft, of course. If my calculation is correct, I get about 700 miles for the distance the two could see each other if the air were clear enough.

IIRC, they were ~1400 miles apart. Something like 7 degrees of latitude and 19 degrees of longitude. I'm trying to find the translated article that includes the Air Comet position, but I can't seem to locate it.

[zipper]

The storm was 400 miles out to sea.

I doubt that in that part of the world there was any reliable radar information.

You might be surprised

http://www.weathergraphics.com/tim/af447/

[zipper]

Even if the wing stalls, the plane just falls. At 35,000 feet this isn't really a big deal.

Actually it's a REALLY big deal. Just the buffet at high TAS approaching a stall can cause delamination of the tail control surfaces on a large transport-category aircraft, which can lead to catastrophic failure.

[genghis]

Clinton lied and got away with it, Zeo0 lies every time his lips move.

i am not a conspiracy fan, but i read an interesting book about flight 800.
There are many unanswered questions.

[old curmudgeon]

The following is from the simulator syllabus for the 747.

---

http://ops.precisionmanuals.com/wiki/PMDG_747-400_Type_Rating_Course_Lesson_3

Stalls are for Horses

The last maneuver we are going to focus on today is aerodynamic stalls.

Lets get one thing clear about stalls in transport category aircraft before we begin: You don’t wait for the airplane to stall. This is not a Cessna, and stalls are extremely dangerous in swept wing jets.

The purpose of this exercise is to teach you to avoid stalling the airplane by showing you the IMMEDIATE actions that you should take upon the first indication of a stall.

What do I mean by the first indication of a stall? Easy: If the airplane begins to buffet as a result of a stall onset (not something you’d see/feel in MSFS….sorry…) or if the airplane gives you any indication that it is approaching a stall such as the stick shaker!

We will focus on the stick shaker for this exercise and you should assume that this is your indication of an impending stall- and is your cue to recover the airplane!

Stall recovery is practiced from a variety of configurations, so I will walk you through a basic stall sequence, and then expect you to try them in each of the configurations described.

The stall sequence works like this:

* Establish level flight at 15000’ MSL.
* Set the configuration for the stall (in this case, 250 knots, flaps up, gear up and wings level)
* Disconnect the auto throttle and set thrust just slightly above idle.
* Use pitch trim to keep the airplane in level flight while you slow toward stalling speed.
* Below 180 knots, do not use any further pitch trim, instead use the yoke to maintain level flight.

At the first indication of a stall (stick shaker!) you should:

* Push the throttles full forward. (If you have the A/T armed, pressing TOGA will help!)
* Adjust your pitch to minimize the loss of altitude. If descending, you should pitch up as much as is required to intermittently activate the stick shaker. Once the aircraft beings to climb, use a stable climb pitch attitude.
* Return to your entry altitude and adjust power to maintain 250 knots.

You do NOT want to “dump the nose” on this airplane like you can do in many light aircraft. If you do so, the airplane will likely commence a descent that could prove catastrophic if you were to stall the airplane without altitude sufficient for recovery!

The proper technique for recovering from a stall in this airplane is to get as much power out of the engines as quickly as you can, while holding pitch stable enough to keep from sinking further until airspeed is sufficient to allow you to climb. Do don’t dump the nose!

This is why your PFD should look like as you commence the stall recovery:

[ScreamingFist]

Actually it's a REALLY big deal. Just the buffet at high TAS approaching a stall can cause delamination of the tail control surfaces on a large transport-category aircraft, which can lead to catastrophic failure

So, is that a structural problem or a software problem? The Flight Control algorithms should never, in a perfect world, allow the pilot to rip the control surfaces off.....

[old curmudgeon]

Yes, you can see one TBZ sitting out there all by itself.

The problem is when they are lined up and you have to pick a spot in the line that you are willing to go through. Now, if there are others behind that spot and you picked a spot that had no cells but lots of rain, then you could blunder into one behind.

It looks good in that brochure advertising a radar, but the real world is not that easy.

Also remember that a really hot storm system is like boiling water in that everything in the pot is in motion. You pick a spot to penetrate and in 25% of the time it takes you to traverse the area, two or three more boil up practically in your face.

It may be easier to pick a spot at 35,000 ft. because the anvil tops have saddles between them, but obviously it has its difficulties or we would not see so many aircraft with hail damage, nose cones broken, passengers with green complexions, etc.

[old curmudgeon]

Obviously the stall is not intentional.

It is caused by loss of control.

Loss of control in that type aircraft is almost always caused by something totally beyond the control of the crew and the equipment.

Turbulence, equipment failure, etc.

A line of severe thunderstorms is not where you would look for that perfect world.

[ScreamingFist]

Loss of control in that type aircraft is almost always caused by something totally beyond the control of the crew and the equipment.

I agree. Thank you for your very informative posts....a pilots perspective is always different from a engineers/technicians. Some design and maintain them..... others actually have to fly them.

[expatpat]

If they were 1400 miles apart (more than DC to Chicago), it seems silly talking about the Comet flight seeing anything related to the Airbus. I thought the two were about 30 mins. apart, which would put it at around 250 miles if they were traveling in more or less the same direction (and 1000 miles if in opposite directions), but the 30 mins. may not be right.

[expatpat]

Yeah, a heavy squall-line is a serious problem, although running is usually an option. However, I used to fly to the Delaware beaches for a number of years, and always had a concern for being caught east of a squall-line and all airfields. I’d be forced to penetrate the storms then (no floats!).

[old curmudgeon]

More from the sim training.

Note that the instructor is very anxious that the aircraft not actually stall, but that recovery is initiated IMMEDIATELY at stickshaker activation.

Note also what he says about rapid descents when cabin pressure is lost, the need to be on auto pilot, etc.

Now all of this is far fetched as applied to this accident and I don’t claim that any of it applies here, but it demonstrates that there are literally dozens of scenarios that could have taken place as a result of storm damage, all of which would have discouraged radio conversation.

So the fact that no one used the radio does not indicate that there was a bomb.

---

Now I want you to practice stalls in each of the following configurations:

Gear Up, Flaps Up. Gear Up, Flaps 10. Gear Down, Flaps 25.

You will notice a dramatic difference in each of these configurations in terms of pitch, and “feel” as the airplane approaches the stall. Remember- the moment you hear the stick shaker- start your recovery!
Lets Decompress A Little

When you feel you have a handle on those stall procedures (boring, aren’t they? Stalls are always my least favorite maneuver during recurrent training) its time for us to practice another maneuver that you will likely never perform except in the simulator during qualification.

For this maneuver, please load PMDG_747_TRC_Lesson3_Part2. (Note: This situation is paused when you load it. I did this because it takes some time for all of the systems in the airplane to initialize, check on their status and commence operation when MSFS loads. Since the sim naturally un-pauses itself before the airplane finishes this process, it can cause some unwanted behaviors in certain autopilot modes. As such, please wait until you see the MCP ALT window display 35000 and the HOLD light illuminate before you un-pause this saved flight. You may need to press the SPD button to activate the auto throttle and use V/S to get the airplane level at 35000 depending upon the speed of your computer.)

You will find yourself at FL350, gear and flaps up cruising at Mach 0.84. For this lesson you are going to demonstrate that you can safely conduct an emergency descent to 14,000 feet in a careful, deliberate manner using the appropriate procedure.

For this exercise, we are going to simulate a cabin decompression emergency, followed by an emergency decent. (Note: we didn’t include a cabin decompression emergency- and since this is training, we are going to pretend that you just lost the cabin…)

At the first indication of a cabin decompression, you will first verify that all three packs are set to NORM, and you will switch the outflow valves to MAN and use the switches to drive the valves closed. (You needn’t do this for this exercise…)

The emergency descent is best flown by the autopilot for safety reasons, not the least of which is that rapid decompression can create debilitating pain and disorientation. It’s best to have the autopilot doing the hard work so that you can focus on monitoring the descent to ensure that the airplane remains in control and the maneuver is effected properly.

[jokyfo]

Dummy! It’s all over the net.

OBVIOUSLY the other pilot also saw lightening. He remarked on this particular flash because it was extraordinary.

http://www.telegraph.co.uk/news/worldnews/europe/france/5444168/Air-France-pilots-battled-for-15-minutes-to-save-doomed-flight-AF-447.html

“These were lent more weight today after a Spanish pilot in the vicinity at the time reported seeing an “intense white flash”.

“Suddenly we saw in the distance a strong and intense flash of white light, followed by a downward, vertical trajectory which broke up into six segments,” the chief pilot of an Air Comet plane from Lima to Madrid told the Spanish newspaper, El Mundo. He has reported his observations to investigators.

Some experts have supported the theory that the plane exploded, given the wide area where debris has been found.”

[zipper]

So, is that a structural problem or a software problem? The Flight Control algorithms should never, in a perfect world, allow the pilot to rip the control surfaces off.....

A good question. We're not talking about ripping off the surfaces from an over-g, we're talking about the airflow separation (buffet) that precedes a full stall. Modern large transport-category jets will most often have low-speed protection (after stick shaker, throttles clamp and increase power, autopilot servo engages to lower pitch) which will prevent a full stall below a certain computer-derived speed, but it doesn't take a full stall to do damage, especially at high true airspeeds. I know of a couple where aircraft didn't develop a full stall but nevertheless had very serious delamination that was undetected by the crew in the air and only found later -- serious enough to classify as an accident by the NTSB.

[justlurking]

Wind shear is not a factor in thunderstorms at altitude. Turbulence (non zero vertical wind speed) is. Even YOU note that these speeds might reach 100 mph. (Normally zero folks!) And, yes, I know TAS is lower at atlitude so the margin over the stall gradually decreases.. So what. Even if the wing stalls, the plane just falls. At 35,000 feet this isn't really a big deal.

OK, step back and take a deep breath -- and read what I wrote, rather than what you think I meant. I said that the speed deviation alarm (sent by ACAS) may have been caused by wind shear. I didn't say that caused the plane to stop flying. I did point out that it caused problems for the Delta flight at DFW, because it was near the ground.

But, let's set that aside for a moment and review what you wrote above. Wind shear is not confined to the horizontal plane. There is also vertical wind shear:

Wind shear refers to the variation of wind over either horizontal or vertical distances. Airplane pilots generally regard significant windshear to be a horizontal change in airspeed of 30 knots (15 m/s) for light aircraft, and near 45 knots (22 m/s) for airliners. Vertical speed changes greater than 4.9 knots (2.5 m/s) also qualify as significant wind shear for aircraft.

In fact, vertical wind shear is a requirement for development of severe thunderstorms:

Thunderstorms in an atmosphere with virtually no vertical wind shear weaken as soon as they send out an outflow boundary in all directions, which then quickly cuts off its inflow of relatively warm, moist air and kills the thunderstorm.

Horizontal wind shear in a thunderstorm is most pronounced at ground level, especially in a microburst. But it also happens at the upper levels of a thunderstorm, as the airflow goes up, cools in the upper atmosphere, then mushrooms out before sinking to the ground to repeat the cycle:

Turbulence is associated with wind shear, but it's chaotic and non-deterministic -- not blowing in one direction (up or down).

Wind shear can be caused by many different things. Air France 447 went down in what is known as the Intertropical Convergence Zone, and the strong thunderstorms forming in that region are a consequence of the meeting of two prevailing atmospheric flows.

You mentioned wind shear as a possible factor. Some people have heard of wind shear and don't really know what it is or why it is important. I put you in this class and called you on it.

I think you are a bit confused about which one of us knows what wind shear really is.

[justlurking]

I am only making the point that as wonderful as airborne radar is, it is not infallible.

It wouldn't have helped the Air France flight, but these days you can get glass cockpit systems in a Cessna 172 (!) that overlay radar images over your flight path.

The images from the national network of ground-based Doppler radars are broadcast over the XM Radio satellites: http://www.xmwxweather.com/.

[old curmudgeon]

Interesting.

I have lost contact with the new stuff.

When I quit, I canceled every single aviation magazine subscription and refused to become one of those airport hanger bums.

I would rather be outside with my horse.

I do know that these ideas were being discussed.

Is the aircraft presented in an accurate manner related to satellite weather so the pilot can actually “thread the needle” where cells are close?

Of course, the ideal would be both that and airborne. Use the satellite weather to get the big picture and choose a better route and the airborne to handle the close encounters.

[justlurking]

A lot has changed. The FAA has been slow in rolling it out to the rest of the country, but a pilot program in Alaska has replaced radar tracking (primary and secondary returns) with "automatic dependent surveillance", a term that basically means that airplanes squawk their positions periodically, using info from an on-board GPS. Ground stations collect the transmissions and integrate them into a coherent view for ATC. Individual aircraft can also listen and plot the location of aircraft near them on an on-board display.

The radar images are only near real-time: I think the lag is up to 10 minutes. So, I wouldn't want to try to use them to thread the needle, without something on-board. But, the ground radar composite addresses some of the problems with attenuation.