AF447

[Broomstick Flier]

Simulator

Hi,

I just received this from a friend who happens to be a capt. on the A320 and flight time on the RHS of the A330 as well. I don't know the source and when/where the simulator flight was done

Dear Ladies and Gentlemen,

The following information that is posted in this reply is something that may interest you all.
The scenario was conducted several times and the results at the end of each scenario produced consistent findings.

In an A330 simulator at FL 350 with a gross weight of 210 tonnes in ISA+10, with icing selected, the aircraft approaches a thunderstorm with a high intensity of turbulence. Due to the extreme turbulence, the autopilot disengages. Shortly thereafter a malfunction is selected to block both captain and first officer's pitot tubes to replicate extreme ice formation.

The airplane reverts to alternate law with protection lost. There is a speed flag on both the captain and fo's PFD. The severe turbulence activates repeated stall warnings. Manual thrust is being used at this time. The speed on the standby altimeter is reading 240kts or thereabouts with MACH .72. From the GPS the ground speed is 350 kts or thereabouts. It is very difficult to read the instruments and ECAM warnings.

Updrafts take the aircraft up to FL 370 and produces a negative G of .2. The aircraft then enters severe downdrafts and the rate of descent averages more than 19,000 fpm. The instinctive reaction is to pull on the stick to arrest the rate of descent. The aircraft shakes and buffets violently. The G force on the SD reads +5 but the instructor's panel shows +8. The aircraft breaks up in flight around 20,000 ft.

After several attempts at this with all results being equal one could not see AF447 sending out any distress signals if this is what happened to them. Applying an unreliable airspeed memory item would have proven to be very difficult because of the violent shaking and opening a QRH for an ADR check procedure even less likely.

Food for thought...

[HarryMann]

What's missing is the "WINSHEAR Prediction system Fault" and GPWS TERR Fault!! what happend to those ?? Because they should be right there along with the other faulty stuff associated....

Could it be those are perhaps inhibited in the cruise ?

[JD-EE]

Broomstick Flier

Is your report perhaps a recycled version of Poison's report, http://www.pprune.org/rumours-news/3...ml#post4987398

[Will Fraser]

ELAC Put on your RayBans and read my hostile question re: AirBus, the a/c that is fully automated, almost flies itself, recommends no one touch the a/p in turbulence, but as it deftly maneuvers through hostile weather, it quits when it can't take the level of weather/dataloss/ etc. and drops it back in the laps of the pilots it generally disdains, but thinks nothing of quitting when the 's' hits N1. This with an automatic penalty in controllability, to wit, cancel roll, no protex, stab trim only, and 8 degrees of Rudder travel, still enough if cycled to rip off the VS. ????

Will

[syseng68k]

"The speed on the standby altimeter is reading 240kts or thereabouts with MACH .72."

Sorry to be picky, but ?...

[ClippedCub]

Updrafts take the aircraft up to FL 370 and produces a negative G of .2

Don't generally think of airlines under negative g, and g of -0.2 means the aircraft went from 1.0g, or more, giving a -1.2g or higher increment. When they developed the F-16 side stick controller, they found the pilot, (arm mass?), over controlled under negative g. To alleviate this human factors, (arm mass dynamic?), they limited forward travel of the side stick controller to a tiny amount, 1/100th of an inch if I remember. All other motions were in the x/16th scale, i.e., 5/16 inch travel.

Don't know if the sim replicates that, but if it doesn't then the 'push' could have been greater, setting up a steeper descent.

[ClippedCub]

One of those forms would be overloading the pilots with "useless" (under the circumstances) information.

In modern fighters, the weapons systems are very complex. The flight control/management computers are trained to only bug the pilot with sequenced, pilot action only, messages during combat engagement, since he's busy targeting, etc.. For example, if he gets an engine hit, the computer extinguishes the fire, shuts down the engine, and tells him he lost one. No pilot action required. Would have thought some of that logic structure would be present in civilian airliners.

[HarryMann]

OK, three pitots fail, (it happens). should the a/p disqualify itself, drop out and demand by SOP that the pilots fly pitch and power (not a memory item, here.

Well a sensible SOP then would be to lookup equiv. attitude/power when first established in new cruise condition? Effectively a refresh memory item.

[Will Fraser]

Harry, it's bumpy, dark, noisy, and the glass is downtown Las Vegas on tilt, let's read from the Bible?? Darn, where is that mfer.

Let me admit at this point to my brethren that I have strayed, I will take a time out.

[ClippedCub]

Both the F-16 and B-2 are statically unstable without the artificial stability provided by computers. (Some of the WWI fighters were unstable too, but the pilot could figure it out because of the low speeds.) The F-16 is unstable so it it can pitch faster in combat. Somebody figured out that it didn't need to be that way years later, what I suspected, and that designing for pitch instability was not necessary since you're at M=0.7 and above. Everybody just got excited about something new since they could do it, and advanced the technology.

[HarryMann]

OK, 'when first establishing new cruise condition'... that'd be way back when life was simple and the lights were on.

BTW. Does anyone know the undamped dutch-roll-yaw amplitudes or frequency?

[bsieker]

Originally Posted by Will Fraser

[...]
This with an automatic penalty in controllability, to wit, cancel roll, no protex, stab trim only, and 8 degrees of Rudder travel, still enough if cycled to rip off the VS. ????

Glad to see PJ2, ELAC and others still taking the time to write long and knowledgeable replies and explanations. Let me add perhaps a few comments, though not quite as eloquently.

Will Fraser, would you please, like so many others, get your facts straight before asking such questions.

Again, what is Alternate Law (Protections lost)?

It is a load-factor demand control in pitch, and roll is controlled as in any conventional aircraft, direct stick-to-surface deflection.

Where does the idea come from that in this control law aileron control was lost, and only manual stabiliser trim and rudder was available? What you are talking about is the last reversion step, still far removed from Alternate Law. It is only in effect in case of loss of all Flight Control Computers.

The rudder travel limiter is independent of Normal or Alternate Law, although it is obviously not independnt of reliable airspeed information. So that Alternate Law and RTL fault co-incide is not surprising here, but Alternate Law does not automatically disable the RTL.


There is seemingly even more confusion about what the computers do, can do, don't and cannot do, both in Airbus and in Boeings.

What people talk about regarding "the computer flies the aircraft" is true for most of the time in most flights in all airliners. The computer doing this is called an Autopilot. Boeing and Airbus are no different.

What the fly-by-wire system does is steer the aircraft precisely and deterministically in the way the pilot commands it to. It rolls at a commanded rate and gives a commanded g-load factor. These are some very basic feedback-control loops, nothing could be farther from "Artificial Intelligence". In no situation is there anything remotely resembling a computer "trying to outsmart the pilot". Whether or not over-reliance on automation leads to an erosion of basic flying skills is another matter, but looking at recent accidents with Airbuses, Boeings and Bombardiers I think that problem is not specific to Airbus crews.

The most prominent exclamation cited to show how confusing computerised aircraft are is the "What's it doing now?" Again, this may be true, but it always relates to mode confusion, which is a problem with modern, highly complex autoflight systems ("Autopilot/Autothrust"). It is common to Airbus, Boeing, Embraer, Bombardier, and has absolutely nothing to do with the fly-by-wire system.


DC-ATE,

During Flight-Law-degradation in Airbus, the pilot does not have to figure out what is going on, and does not need to remember how to control it entirely differently in Normal, Alternate (red. prot), Alternate (no prot) or Direct Law. It's always the same: push the stick: nose goes down, pull the stick: nose comes up, push it left: plane banks left, push it right: plane banks right. It is an airplane, and it is flown like an airplane. The difference is the level of protection (which in this form, non-FBW aircraft don't have in the first place), and the amount of "smoothing" the computers do to achieve the desired trajectory.

Even if it crops up again and again in this thread, there is absolutely no indication that flight control had reverted to manual reversion, i. e. rudder, pitch trim and thrust only.

Originally Posted by WhyIsThereAir

but the software will not let the plane be flown until the warnings are cleared.

Where does this come from? Are you seriously implying that, like a modal dialog in Windows Software, all operation of the flight controls is blocked or suspended until all ECAM procedures have been finished?

The aircraft can always be flown and will always be flown if that is necessary before any troubleshooting and other procedures can begin.

Again, that this is not always the case is well-known, but by no means an Airbus-only problem. An Adam Air B737 was lost not too long ago, because both pilots were troubleshooting the ADIRUs, and no-one was flying the airplane. It broke up in mid-air when recovery was finally attempted after the bank angle had reached 100 degrees, because in trying to arrest the steep dive the ultimate load limit was exceeded. Even in Alternate Law, Airbus would have prevented excessive load; if I remember correctly, the Adam Air flight reached some 3.7G before it disintegrated.

It is interesting to note that this flight broke up because it had no load limiting protection, and in the A320 accident at Perpignan the crwe might have wanted a bit more than the limited 2.5G. Whether or not 3G would have been enough to stop the descend before the water surface I cannot say.


As to flying "Pitch & Power": As has been mentioned before, the memory items are: disocnnect A/P and A/THR, set CL thrust (will be set as soon as A/THR is disconnected in the "normal" way), and 5 degrees pitch up (depending on altitude and flight phase).

There are then longish tables to be consulted which exact N1 value and pitch angle is to be set at which altitude. This may indeed be nigh impossible in severe turbulence at night.


ClippedCub,

Sometimes an engine that is on fire is still producing thrust that may be needed. It is the pilot's decision (on an airliner usually not engaged in a dogfight ...) when to shut it down. It should not be left to the computer. And on an airliner flight deck there are two pilots, unlike most modern fighters.



Bernd

[ClippedCub]

Engine fire point taken. Just an example. If the software is saturating the crew at the worse time with warnings and paper look-up requirments, perhaps some priority setting criteria would be useful. The military house doesn't always talk to the civil side.

[ELAC]

ELAC Put on your RayBans and read my hostile question re: AirBus, the a/c that is fully automated, almost flies itself, recommends no one touch the a/p in turbulence, but as it deftly maneuvers through hostile weather, it quits when it can't take the level of weather/dataloss/ etc. and drops it back in the laps of the pilots it generally disdains, but thinks nothing of quitting when the 's' hits N1.

Will,

You are so seriously barking up the wrong tree that it's hard to know where to begin. For starters I'd have to guess that you've never flown any other jet transport aircraft with an autopilot then? If you did you'd know that each of them has limits to their range of operation after which they must disconnect.

For example, ever try doing a go-around on the autopilot in a 757 or 767 with an engine failed? If you did you'd know that there was always a point where the autoflight went from triple to single channel at which you could expect the autopilot to kick-off due to insufficient remaining rudder control at certain trims, weights and configurations. By contrast I never once witnessed a similar event in an A320/A330/A340.

Any autoflight system can only work within its designed and certified limitations and when those limitations are exceeded the autoflight system must return control to the pilot. From my experience with both Boeing and Airbus the usable range of the Airbus autopilot is every bit as broad, and probably broader than that of any other jet transport in current use.

In a situation as extreme as the one you suggest (which is simply a hypothesis, hardly a certainty) pretty much any autopilot would have packed it in. So, if that's what occurred in this case, it had very little to do with the manufacturer and a whole lot to do with the extreme nature of the conditions encountered.

This with an automatic penalty in controllability, to wit, cancel roll, no protex, stab trim only, and 8 degrees of Rudder travel, still enough if cycled to rip off the VS. ????

You really have no idea what you're talking about do you? You are mixing up flight control laws and autoflight, two entirely different matters. Further it seems that you believe that all flight control computation was lost ("stab trim only") when there isn't even the slightest evidence suggesting that. The only thing we know with respect to the aircraft's flight control laws is that they changed to Alternate Law (Protections Lost), which in itself is hardly a crisis. I should know having flown and landed Airbuses in Alternate and Direct law.

Regarding the rudder on the A330, rudder deflection is limited as a function of speed. After a rudder travel limiter fault "the maximum rudder deflection remains at the value reached before failure". That value is dependent on speed, which according to my manuals at M.80 and FL350 (272 KIAS) is approx. 6 degrees (not 8). By contrast, if a similar failure (Rudder Ratio) occured on a B757, do you know what you'd get? The answer is that on the B757 the only thing that happens is that 1 of 3 hydraulic actuators is depowered and the pilot is given a warning to "Use rudder with care above 160 kts." So, you tell me, which rudder control system would provide the greater margin of safety in the conditions you described? One that was continuously calculating a safe limit and imposed it at failure of the normal system or one that simply reduced rudder control power by a flat 1/3rd and said "Take Care!"

The point of all this, Will, is that you are chasing after your own personal bogeyman without knowing in detail how the A330 or other similar aircraft would have handled the scenario which you suppose may have caused the accident. You have no facts that support your supposition so instead you start prevailing on PJ2 or others to provide you with what you consider a smoking gun, but which anyone who is professionally familiar with the aircraft would tell you is either inconsequential or inconclusive.

ELAC

[DC-ATE]

bsieker -

The difference is the level of protection (which in this form, non-FBW aircraft don't have in the first place), and the amount of "smoothing" the computers do to achieve the desired trajectory.

Ah.....but my "old" airplane, along with others with 'conventional' controls, HAD this protection: it's called Design Maneuvering Speed; the speed at which application of FULL available aileron, rudder or elevator will NOT overstress the airplane. It approximates Rough Air Speed.

[ELAC]

Quote: DC-ATE
Ah.....but my "old" airplane, along with others with 'conventional' controls, HAD this protection: it's called Design Maneuvering Speed; the speed at which application of FULL available aileron, rudder or elevator will NOT overstress the airplane. It approximates Rough Air Speed.

And you figure that's not available with "new" airplanes? Enough said and my point made about why this thread has degraded to catering mostly to the rants of the uninformed.

ELAC

[EmBee]

When should the a/p disconnect? I guess that is less about raw technology as the nerve of manufacturers safety engineers, pilots unions or the press (pilot error, particularly in difficult circumstances, is usually graciously forgiven, computers never!). For a peep at the future of the a/p and FBW try:-

Rockwell Shows Off Self-Healing UAV | AVIATION WEEK

or

Flying on a Wing and a Half - Defense News

…. although I guess that software would have had a hard time without attitude data or perhaps airspeed data!

[selfin]

Originally Posted by DC-ATE

Ah.....but my "old" airplane, along with others with 'conventional' controls, HAD this
protection: it's called Design Maneuvering Speed; the speed at which application of
FULL available aileron, rudder or elevator will NOT overstress the airplane. It
approximates Rough Air Speed.

There is a widespread misunderstanding among pilots about the degree of
structural protection that exists when full or abrupt flight control inputs are
made at airspeeds below the maneuvering speed.

NTSB/AAR-04/04, p. 160, Conclusion #16.

[BarbiesBoyfriend]

Basic hand flying skills must be more difficult to maintain on the Airbus series than on any other type.

When it all goes 'tits up', ANY pilot should be able to fall back on the basic skill-set he learned before his first solo.

We all understand the basic utility of "automatics".

They are there to allow you to eat your breakfast or read the paper or look out the window (!).

They are NOT a substitute for that stuff you learned at your flight school!

I see young FOs these days that can barely fly the a/c.

The real worry tho' is that they think they are completely 'bulletproof'.

I'll bet a quid that this a/c was being flown by the F/Os.

If I'm right, and they were typical 'autos ON' F/Os, they might have found their lack of recency in holding the dog**** controller less than helpful.

[Chu Chu]

I always figure the simplest explanation's the most likely, and the simplest explanation of the VS breaking off is impact with the water. But if it came off early on, could the resutling yaw have caused the inconsistent pitot readings?