Air Asia Indonesia Lost Contact from Surabaya to Singapore

[silverstrata]

Ben There:

Setting 90% N1, on the A320, will assure a safe airspeed (assuming you're not approaching a stall)

An even better strategy, is to select 80% and a degree or two of attitude below normal cruise.

What you want to do is get out of coffin corner, especially if things are getting turbulent, and by setting a lower thrust and attitude you will do just that.

Manaada:

I don't fly Airbus, but I find it a bit strange that an aircraft with so many protections built in, will allow the autotrim to trim the stab into a position that you simply can't fly out of.

So will the Boeing, as was amply demonstrated on an approach into a European base a few years back, when the pitch got out of control.

Unfortunately, there are still some corners of the envelope that will always remain outside controllability. You want a trimmer, but that trimmer can also reduce elevator authority (as it can on the Boeing). You want lots of thrust, but that thrust can reduce elevator authority (as it can on the Boeing). You want rudders, but that yaw can diminish roll control (as it can on the Boeing).

With an aircraft, both in design and in flight, you are juggling any number of interconnected and competing controls and parameters - any or all of which can be present at the same time. And from that melee of actions and reactions you need to create stable flight, which is not always possible at the far edges of the envelope. So the idea is (and always has been) to remain within the established safe envelope, where all of those many competing aeronautical parameters can coexist and operate in safe harmony.

The problem lately, is too many aircraft exploring the edges of the envelope where they should not even dream of going. So rather than fiddling with the edges - changing this law or that law - we need to concentrate more on not letting aircraft get to the edges and corners of the envelope where they have absolutely no business being.

One minor mitigation would be to take fuel restrictions off pilots, so they are not flying in the higher corners (and adjust the pax figures to compensate). But no company will do that....

[mtwittm]

Law

Maybe we need Lawyers instead of Pilots to fly these airbus aircraft in order to interpret - Direct Law, Alternate Law...etc.

[Leightman 957]

Trim

Nigel@2677: "An Airbus is slightly different.............it is not "trimming" as such, but aligning THS and Elevator to retain full elevator control."

But "full Elevator control" at an AOA (and airspeed, temperature, pressure etc) presumed by the computer to be correct, along with previous pilot inputs. I thought that was exactly what happened in the case where excessive (and unnoticed?) FCC-directed the THS trim to alter to the point where elevator authority was insufficient.

If the above is incorrect, are you saying that the FCC will correctly notice and interpret unusual variations/combinations in AOA, OAT, pressure, and airspeed over a time period that varies from nearly instantaneous to several seconds, and during that time correctly interpret that a sensor set is beginning to disagree? If you are, the system is pretty smart. But if the system can't interpret all those inputs and correctly categorize them as this or that type of situation where action A is the best, the pilot has lost 5 or 10 or 20 seconds while the computer thought, then annunciated, and then the pilot began to redirect his attention from wherever it was.

I'm getting back to the time involved in the FCC process here. Explanations by posters about the FCC logic chain haven't mentioned time. The system must ignore (or filter) instantaneous changes as either transitory or anomolous, so that means that a certain number of samples over time must be taken for all sensory inputs. Can someone knowledgeable speak to the amount(s) of time the system consumes before it acts in various ways?

RE AOA: Because AOA failure seems to have recurred, do not (or could not) the sensor be mechanically exercised to confirm its movement is free? This need not necessarily involve actual movement but could be forced applied to a pressure transducer. Or is this already built in?

[Rananim]

One of the first things I was taught in primary training is, "Don't trim into a stall." Apparently, the AB system will do just that in some circumstances

Autotrim when handflying is bad news; nice and geeky when all is well but precisely what you don't want when things go pear-shaped. Either you're fully flying it ie trimming as well or the AP is in.

100% correct.Strange counter-intuitive design.I dont like it.And I dont need to fly it to know that.In normal flight,pilot doesnt trim so becomes accustomed to not trimming.When something happens,the computer has trimmed into the stall and then the computer leaves it to the pilot to manually trim.No,its not right.No pilot would ask for that design.Keep it simple.Human brain doesnt work well with startle factor and lots of ECAM warnings blaring.

[Coagie]

I think how QZ8501 and AF447 got into stalls and attempted to recover will be different. The pilot flying in the AF447 crash, Bonin, makes it difficult to compare it to other air accidents. Reading the CVR transcript, makes it clear to me, even through translation from French to English, that Bonin was very uncomfortable, even frightened, well before the UAS incident even took place. He was spooked (frightened) by his first exposure to St Elmo's Fire and the the "electrical smell" AF447 encountered. Most, especially when the phenomena were explained to them, would think "That's cool" or "neat", but Bonin reacted as if they were supernatural, as if he'd seen a ghost. He may have even known something about what to do in a stall situation (wasn't one of his hobbies flying gliders?), but he was already in "fight or flight" mode (wrong kind of "flight" mode. No pun intended), and using his reptilian brain instead of his reasoning brain. He didn't think things through or go back to his training. Each person is different, but Bonin "cracked up". Some people would use the term, "freaked out". The threshold to "crack up" is different for everyone, but can change with training and experience. This is why some training/testing involves putting the student in pressure (stressful) situations, to see if they can still perform. Some people perform and pass testing fine, when not under pressure, only to falter, when pressure is applied. Military training is often known for this, but so are some civilian training regimens. I think the amount of "Pressure Testing" and "Trickery Testing" in airline pilot training and testing, needs reviewing and revising to better insure, that a pilot's "cracking up" threshold isn't too low. Some schools and airlines may already have it "right" and some may have it "wrong" or some degree in between.
Bonin had a low "cracking up" threshold in the situation he found himself in, and since every person can be vastly different, it may be apples and oranges to compare his incident and the Air Asia QZ8501 incident, as far as how they got in a stall, and how they attempted to get out of it. I think his "cracking up" threshold just happened to be lower than an average person, and, I imagine, much lower than an average pilot. I can't bring myself to blame the computer in the AF447 crash, as it didn't throw an unmanageable situation in Bonin's lap. He created the unmanageable situation. It could prove different in QZ8501
Anyway, hopefully, we'll see, when an Air Asia QZ8501 CVR transcript is published. I don't think the chances are that another pilot, with a low cracking up threshold, slipped through the cracks and caused a crash in the same way. Hopefully, I'm not wrong, or things are worse than I thought.

[Coagie]

"Coagie

You aren't suggesting actual, all round air experience might even count towards the selection criterion now?? Goodness gracious me, what are you on?"

Sop_monkey,
It's the madness flaring up. Forgive me.

[glendalegoon]

CNN reporting 2 airasia bodies found 600 miles away

cnn just reporting and will comment in minutes about finding 2 bodies 600 mies from crash site.

[NigelOnDraft]

One of the first things I was taught in primary training is, "Don't trim into a stall." Apparently, the AB system will do just that in some circumstances

Having reviewed the FCOM in fact the auto-trim stops as AoA protection kicks in i.e. well prior the stall (providing AoA protection is working of course...)

[Neptunus Rex]

CNN reporting 2 airasia bodies found 600 miles away

Without flotation assistance, such as an inflated life vest, a human body immersed in water will sink. It then takes at least 48 hours for the decomposing body to produce enough internal gases for the body to rise to the surface. During this process, as well as afterwards, the body is subject to any local currents.

As the accident occurred on 28 DEC 14 and these two bodies were found on 27 JAN 15, 29 days have passed, which amounts to an average movement of under one knot for the whole period.

"Do the Maths."

Perfectly explicable.

[enola-gay]

Thread Comment

Regular as clockwork, someone pops up to say" Wait for the official report. Stop speculating"

Aye just like the Chilcott inquiry!

If you read the blog title it is called "Rumours and News", so I presume people are allowed to post published news, speculate about its veracity and consider all the rumours that abound?

If not change the blog title to "Facts Only"

[mickjoebill]

Numerous posters have pointed out that from their first hand experience there is a significant difference between severe turbulence in real life compared to the comfy environment of the sim.

Should task specific "turbulance and stall sims" be used?
They need only have essential cockpit instruments to reduce building and running costs. One version for bus and a second for boeing.

This bare bones, one size fits all generic cockpit concept has proven successful for HUET training.

Is there enough data from recent stall spins or severe turbulance to replicate the physical environment?

Rather than a full size cockpit on multiple actuators a smaller capsule on the end of a robotic arm is able to generate the G forces in any plane, like these F1 and Diamond aircraft sims.
German researchers claim first: Robotic flight simulator - AOPA

Giant robot arm used as F1 simulator - CNET

[AirScotia]

Numerous posters have pointed out that from their first hand experience there is a significant difference between severe turbulence in real life compared to the comfy environment of the sim.

Many pilots here have reported how difficult it is to see the instruments in severe turbulence conditions. Do simulators adequately replicate this effect?

[Machinbird]

FAC Computers Getting the Eye

As mentioned yesterday, the FAC computers are being considered as a possible causative element in the QZ8501 accident. Leaked information from those who do not wish to be identified has indicated this area of interest by the investigators. QZ8501: Probe team vets possible computer glitch, crew response | New Straits Times

"The aircraft remains fully controllable if you lose the two FACs,” an Airbus spokesman said by email.

This may be dissembling however. Suppose the FACs remained engaged in Normal Law and failed to detect a failure of that system. Could the FAC system then activate overspeed protection without an actual overspeed?

This need not be a design issue, it could also be a maintenance issue. In any case, if the DFDR data continues to point to the FACs, they should probably salvage the cockpit and recover those boxes for examination.

[A320FOX]

FACs

The FAC "boxes" are not in the cockpit, they are under the forward galley in the E & E compartment.

[scard08]

(It would also seem that Airbus now has at least two sets of pretty good data on entry to stall, stall behavior, transition to deep stall, flat spin, 10,000+ per minute altitude loss.... to add to their simulator's knowledge base.)

I was thinking about this too. Does the FDR from a crash that involves a stall help improve the simulated responses to a stall?

The answer is obviously yes, but how many sets of data about performance outside the envelope (i.e. flights) do you need before you can program a simulator to correctly simulate performance there? What does it take to make simulations significantly better? My gut feeling is "a lot".

Any folks who build simulators out there? Any test pilots who can speak to this?

[HeavyMetallist]

Originally Posted by Machinbird

This may be dissembling however. Suppose the FACs remained engaged in Normal Law and failed to detect a failure of that system. Could the FAC system then activate overspeed protection without an actual overspeed?

This need not be a design issue, it could also be a maintenance issue. In any case, if the DFDR data continues to point to the FACs, they should probably salvage the cockpit and recover those boxes for examination.

Apart from the fact that the FACs aren't where you think they are, I think you'll also find that the ELACs and SECs are the flight control computers that apply the control laws, and not the FACs.

[Propduffer]

News from the data and voice recorders is apparently leaking out.

The WSJ has a story including this:
They have said icing of speed sensors, a phenomenon that contributed to numerous fatal airliner crashes and serious incidents over the years, doesn’t appear to have been a factor this time.
AirAsia Co-Pilot Was Likely at Helm Before Crash - WSJ

[RifRaf3]

Notice in the video that it's rotating at about 1 rev/sec to simulate G. If you are pulling straight out of a dive you are not rolling at 360 degrees/sec, so you gain one simulation aspect (sustained G) at the expense of another, namely rotation on another axis. Only a very large, 3D printer style, framework could do a reasonable job of isolating each axis of rotation and the cost would be prohibitive.

[Machinbird]

Quote:
Originally Posted by Machinbird
Suppose the FACs remained engaged in Normal Law and failed to detect a failure of that system. Could the FAC system then activate overspeed protection without an actual overspeed?

This need not be a design issue, it could also be a maintenance issue. In any case, if the DFDR data continues to point to the FACs, they should probably salvage the cockpit and recover those boxes for examination.

Apart from the fact that the FACs aren't where you think they are, I think you'll also find that the ELACs and SECs are the flight control computers that apply the control laws, and not the FACs.

HeavyMetallist, I'll have to defer to your experience on the actual physical location of the FAC computers.
Guess they will have to grab the piece from ahead of the wing (which may also include the cockpit).

As to the exact manner the FACs interface with the ELACs and SECs, my reading of FCOM indicates that it is the FACs that have the function of monitoring the flight envelope and invoking the protections (among other functions). The ELACs and SECs would then implement the response. Take a look at FCOM 1.22.40 Auto Flight-Flight Augmentation-and see if you read it that way.

[B4MJ]

Has anyone ever explained why AF447 PF obsessively pulled back?

Survey some old experienced simulator instructors. Ask how many times they've observed pilots pull back (even obsessively) on the yoke/stick when they're suddenly startled (frightened) by an upset.

Human nature isn't going to change and until upset training becomes really realistic (and everyone is forced to receive it), this type of event will always be lurking.