Air Asia Indonesia Lost Contact from Surabaya to Singapore

[mseyfang]

I'll take the computer.

Here's the thing about a computer--it still has enormous upside because the software is the result of a collective endeavor. It is true enough that in short run there might be more accidents but in the long run it will produce significantly less accidents than any human. We have seen this is every field that computers have been involved in.

It is not a question of IF the computer will take over for the human pilot, only a question of WHEN.

I don't see why this has to be an either/or proposition. There are times when a computer is better suited to a particular task and times when a human pilot with full control of the aircraft is better suited to a set of tasks that fall outside what the programmers envisioned. Humans make mistakes, but they also are capable of doing things right when things really hit the fan. There is no computer on Earth capable of what doing what Al Haynes and crew did with UA 232 or what that Aloha crew did with their 737 when the roof came off. On the other side, TCAS, windshear detection and GPWS have been quantum leaps forward in making aviation safer, though those systems still require a human pilot to intervene.

I agree with you that the computers will get better, but that still does not eliminate the need for sound airmanship in commercial aviation, particularly in the area of judgment. Computers still do as instructed and as the saying goes, garbage in gets you garbage out, as the 1995 Cali accident demonstrated when the wrong waypoint was entered into the FMC. We may get to the point where once the plane is pushed back it starts, taxis and takes itself off, then lands and taxis itself to the gate. Such an aircraft will also fly merrily along its pre-programmed course into a thunderstorm or other peril without human intervention.

I'm not against technology, but I also think that pilots ought to have the skill set to safely plan a flight and hand fly if conditions warrant and I don't think that's asking too much because no computer can be programmed for every contingency that may occur.

[RatherBeFlying]

The computers do precisely what the systems analyst and software designer wanted them to do...

The computer programs purport to do what is wanted.

In truth, the computer programs do precisely what is coded which is not necessarily what was wanted.

Sometimes the programmer does not correctly program the design.

Sometimes the design falls short of what is required.

One very good way to sabotage a project is to program exactly what is specified

All the above applies to de novo projects.

These days the majority of projects are based on interfacing to older code that has been lurking about for several years. Making old code do new tricks can expose you to gotchas and shortcuts that nobody remembers any more. You have to enter the mindset of programming decisions made and forgotten several years ago

Understanding old code is orders of magnitude harder than writing new code

[Methersgate]

Dr Philippa - re2986: The use of three inputs is a concept that is older than aviation. Ships carried three chronometers to determine longitude, because if one malfunctions this can be determined by reference to the other two. Two are half as safe as one; three are twice as safe as one.

[silverstrata]

Gendoon:

well, in certain scenarios with the 737, that is exactly what you do, you turn off the hydraulics esp to the rudder.

Only because some twit made a single actuator rudder, and an even bigger twit certified it. You cannot hope to design a systems protocol that is founded upon foolishness. And besides, you never cut all the hydraulics to a 737 rudder, as the thing will fall out of the sky if you do. (No cable back-up, I'm afraid).

mountainbear:

What still remains to be understood is why some pilots revert to an incorrect cognitive map under stress.

Indeed, but it happens. I think it is mental overload, when reality does not equate to what the mind thinks it should be. Have seen this many times in ab-initio training, when the aircraft goes beyond the comprehension and capabilities of the student, and they go into tunnel-vision mode or freeze mode.

The only way through that comprehension barrier is repeated training, so that the unimaginable becomes normal and routine. This is called 'training'. But the sim is not so good at this, because you need the adrenaline of a real situation to get the brain steaming. This is why I like computers - they are all the same, they are good to go from day one, they don't pump adrenaline, and they very rarely 'steam'.

[Coagie]

The systems analyst, software designer and acceptance engineers often ask: "What did they do that for? Who would have thought a pilot would do that?

Ian W, I knew a programmer who worked for a company making smartcard compatible washing machines for laundromats. He was the go to guy, that the company sent software to, in order to get the bugs out. He told me the biggest problem, was just as you say. For instance: His fellow programmers would ask questions like "Why would anyone open the door, before the cycle ended? That wouldn't make sense." And they wrote the software accordingly.
Go ahead and convince yourselves, if you want, that there is not a programmer of some obscure, hardly used subroutine in flight control software, that didn't use that line of thought while writing it, since it's more important or more glamorous than washing machine software.
Thankfully, most of these bugs are uncovered with thorough testing.

[glendalegoon]

DEAR silverstrata

A 737 will not fall out of the sky if you cut all the hydraulics to the rudder.

I have a small, laminated type card that says I know stuff about the 737.

YOUR KNOWLEDGE of the 737 and certainly other types of airplanes is such that ANYTHING you say about airplanes is suspect.

Will any other 737 type rated pilots confirm that if the rudder loses all hydraulics you can still fly?

Will any other 737 type rated pilots confirm that turning off hydraulics to the rudder is part of the rudder hardover recovery?

SILVERSTRATA do you care to retract your statements about airplanes?

Will anyone else support my statements? esp 737 pilots?

for quite awhile, older 737s had memory items/QRH which included cutting the hydraulics to the rudder in case of hardover...NG models and modified models have additional protections and systems.

But dear strata, if you can't move the rudder, you can still get down in one piece

[DrPhillipa]

And if two pitot tubes read 0 kts and one reads 350 kts what is the speed?

If you have WOW probably 0 and one binnable pitot, at FL 300+, 2 are probably frozen and the other is probably nearer the truth but possibly also a bit iced and therefore not 99.95% reliably accurate.

So what to do. 9 times out of 10 the AP could probably safely fly pitch and thrust and descend to below FL350 based on more relaxed numbers, machine experience and probabilities. Maybe 99 times out of 100. Could even broadcast a Pan Pan for you and check for collisions.

But!

[Metro man]

The CBs for FAC 1 are on the overhead panel and can be reached from the captains seat, the CBs for FAC 2 are on the rear panel and would require the captain to leave his seat.

If for some reason the CBs needed resetting then why didn't he take over control himself and instruct the F/O to perform the reset ?

I have no hesitation in taking over if in my opinion the situation warrants it. Once the unsatisfactory aircraft state is resolved control can then be passed back at a appropriate time.

[MountainBear]

I don't see why this has to be an either/or proposition. There are times when a computer is better suited to a particular task and times when a human pilot with full control of the aircraft is better suited to a set of tasks that fall outside what the programmers envisioned. Humans make mistakes, but they also are capable of doing things right when things really hit the fan. There is no computer on Earth capable of what doing what Al Haynes and crew did with UA 232 or what that Aloha crew did with their 737 when the roof came off. On the other side, TCAS, windshear detection and GPWS have been quantum leaps forward in making aviation safer, though those systems still require a human pilot to intervene.

I'd like to address this point because it is based upon what is a common yet faulty understanding of the nature of computer systems software. All software design is a function of collective experience. This collective experience can come through user studies, trial and error, or even by tombstone. It remains a repository of collective experience.

Experience is not infinite. While there are indeed millions of possible permutations at the interface of weather, the physics of flight, passenger comfort, and regulatory frameworks these combinations are finite. Moreover, most of them are not deadly. There is only a small subset, perhaps in the hundreds of thousands, that result in a plane crashing. The result is that the collective experience as codified in software will always and inevitably surpass that of a single human mind. There is only one possible exception to that rule and that is if the person under discussion is omniscient and omnipotent--that is God. No human pilot is God.

So while it is true that that there remains limitations to computer software today evey accident just moves the needle one more degree closer to full. Eventually every possible permutations will be coded for. Indeed, there is a cogent case that right now that human pilots cause more crashes than they prevent. But whether that is true this exact second in time is irrelevant to the fact that the day is approaching.

[SAMPUBLIUS]

for glendalegoon and others

This link should help to explain to the non- pilots about FBW misconceptions, etc. AS to specifics re 737- there should be somewhere on the boeing site the details of 737 systems which mostly have cable backup for at least minimal control.

How does the fly-by-wire in Boeing 777 differs from the Airbus system?

Gives a fair history of FBW and hydraulics and cables from a pilot . .

[HeavyMetallist]

Originally Posted by Coagie

Ian W, I knew a programmer who worked for a company making smartcard compatible washing machines for laundromats.

That has nothing whatsoever to do with how safety-critical software is written for applications such as airliner flight control software, let alone how it's reviewed, approved and certificated. For all the whining here from (apparently) pilots about how many posters are clearly not pilots and therefore haven't a clue about flying these aircraft, it's equally clear that very few of those supposed pilots have a a clue about how their aircraft are designed and engineered. It isn't at the whim of some juvenile nerd, it is thoroughly reviewed and questioned by older and wiser heads at every stage of the design and development process, and there is very direct (and forceful) involvement of professional pilots at every stage, starting from the original design concepts.

[gulfairs]

pitot

I am intrigued, why would the pitot heat fail because of CB activity.

Erroneous IAS is more likely if the static vent became iced up.

Geriatric Flatulance,

[autoflight]

This is a fascinating psychological topic. The general consensus in the literature is that pilots revert to a primitive concept that equates flying with climbing. This makes sense in that in order to fly one must climb up from the ground. What do you do when you want to fly? You go up. Mentally, the pilot goes back to square one and tries to take off all over again. In the AF447 case it was revealed that all the way into the sea the pilots kept trying to get the plane to /climb/ and even verbalized that intent. When in fact is they have just flown the plane they would have been OK.

What still remains to be understood is why some pilots revert to an incorrect cognitive map under stress. Is it that they were poorly trained to begin with? Or perhaps they have some basic psychological predisposition?

The physical, electronic and environmental problems encountered by a crew can occasionally be very difficult. Almost always, a good outcome is determined by the actions of the crew. Psychology cannot be ignored when reviewing why an accident occurred or was prevented, and I am very pleased that this has been raised.

This A320 accident seems related to conditioning and acceptance of known problems, i.e. FAC and weather. No-one was going to rock the boat. Right there is the psychology link.

We are conditioned to avoid severe weather either by flying around it or choice of a flight path through an area less affected.

When there is real difficulty with either of these two options on line-up with a take-off clearance, a slot time and a line of aircraft on finals, a significant number of crews will avoid the torturous prospect of taxiing to an exit to join a long queue again for the next take-off opportunity.

When the real difficulty is en-route, return to departure airfield or diversion to an airfield that may have reduced handling capacity is less likely to be in the DNA of a crew.

When the real weather difficulty is at destination, I believe almost all crews are conditioned for holding or proceeding to an alternate, though some require the prompting of ATC to avoid a hazardous situation.

How do crews press on in situations where appropriate vigilance would dictate more prudent decisions? Common to all three scenarios is disruption to schedule, increased airline costs, insufficient fuel for the particular circumstances and insufficient duty time remaining.
TAKE-OFF:
Miss slot time
Cause missed approaches
Written report required when other company aircraft accept take-off
ENROUTE:
Psychologically conditioned to continue to destination
Embarrassed to return or divert when other company aircraft continue
Never operated to potential diversion airfields and perhaps insufficient knowledge about ground handling and refuelling facilities
Diversion airfield is in another country
DESTINATION:
Other aircraft are landing without any reports of problems
ATC has not closed airfield
Did not previously consider diverting to alternate, updated forecast and TAF not requested. Maybe the alternate weather is the same?

A task for the vigilant crew is to identify and minimise risk, while attempting to operate aircraft efficiently. Good abilities, knowledge of aircraft and procedures is assumed. Here are some examples of other ways to minimise risk:
Know flight and duty time available before departure. Don't automatically consider that your company has taken this into account
Consider fuel and carry more than the flight plan minimum if prudent.
Depart a few minutes early whenever possible. Helps with duty time, slot times and might enable an early or on-time arrival
Get a print-out of forecast and TAF / METAR for all possible diversion airfields at flight planning stage. Do this again via ACARS enroute
Have ready information re diversion airfield minimas, facilities, grid MORA and 25 NM safety altitudes. Make your own map with this info for common routes if company allows. Of course keep it updated.
Depending on planned LW, save a little fuel enroute by direct tracking and good choice of flight level
Figure out the best way to get a good slot and keep it. Avoiding a 1 hour delay helps reduce the pressure to press on.

There are sure to be some experienced crews who give proper consideration to many of my points. From my own many critical observations there are many who will allow their options for safer flight to just slip away without really knowing that it is happening. I am not being ultra precious about psychology. Insufficient thought processes cause fatal events.

I have remained amazed at AF, BA and other major airlines landing at BKK in a severe TS with very heavy rain, strong and variable gusty wind with driving rain. Nobody wanted to take-off, but landing was OK??? At the time BKK tower did not have the authority to close the airfield for adverse weather. The only reasonable conclusion I could draw was that these major operators didn't know this and trusted the open status of the airfield to indicate that conditions were safe when they clearly were not. I have tried to figure out the mental processes that allow such risky behaviour.

[Coagie]

I am intrigued, why would the pitot heat fail because of CB activity.

The heaters don't fail. The pitots ice up, but the heater melts them usually within a minute. They get wet because of a storm at high altitude (usually the air is dry up there, but the storm pushes the moisture up), the plane then goes through very dry air, and the moisture ices up due rapid evaporation (dry enough to pull the heat away, causing the ice, but not enough to get rid of enough of the moisture to prevent the ice). Subtle changes in the shape of the pitot or changes in it's heater, may be the reason why one brand of pitot may work better than the other. There may be other scenarios for pitots with working heaters to ice up. That's just one I can think of. The heaters, in some cases, may cause the icing, because, when it's ultra cold, Ice particles act more like sand (ask arctic explorers), and don't stick together (sleds don't even slide), but the heaters may melt them enough to stick to one another. Sorry the explanation is a bit rambling, speculative, or down right inaccurate. The beer is kicking in.

[Ian W]

Originally Posted by HeavyMetallist

That has nothing whatsoever to do with how safety-critical software is written for applications such as airliner flight control software, let alone how it's reviewed, approved and certificated. For all the whining here from (apparently) pilots about how many posters are clearly not pilots and therefore haven't a clue about flying these aircraft, it's equally clear that very few of those supposed pilots have a a clue about how their aircraft are designed and engineered. It isn't at the whim of some juvenile nerd, it is thoroughly reviewed and questioned by older and wiser heads at every stage of the design and development process, and there is very direct (and forceful) involvement of professional pilots at every stage, starting from the original design concepts.

I have worked on such safety systems with really professional programmers and system designers it would take around 18 months to approve a design in the software through the control process. Complete regression testing was carried out every month on monthly 'builds' and after a year of updates the system went for another year of extensive testing before release for operational use. Quite often within hours faults were found by users.

There is no such thing as fault free software. Some software especially the FGMC software is built to default to fail, or gracefully degrade, rather than be 'clever' as the number of variables and unknowns the system has to work with are too high for safe coding. There will always be these times. Remember, the system has to be built to cope with not all likely inputs, not even all potential inputs but all possible inputs - and that starts to get extremely difficult and costly. It is easier and usually safer and more effective, for the system to pass the bag of bolts to the flight crew.

[MountainBear]

Remember, the system has to be built to cope with not all likely inputs, not even all potential inputs but all possible inputs - and that starts to get extremely difficult and costly. It is easier and usually safer and more effective, for the system to pass the bag of bolts to the flight crew.

That is a justification to pass the buck and not a technical argument. One of the more troublesome aspects of making the computer the pilot is the issue of accountability. Right now "pilot error" is an easy out for the accident investigator. Take away the pilot and then who is to blame---software engineers? The software designer? A coder? Do these professionals now need to carry liability insurance?

But if one's first concern is safety then for every hero captain one has to balance the goats. The question isn't whether the computer will make errors--it will--the question is whether overall the software will make less errors than humans. And that is true or will soon be true.

[_Phoenix]

The plane climbed from 32,000 feet to about 37,000 feet in 30 seconds before entering stall.
The stall warning alarm sounded for four minutes before AirAsia Flight QZ8501 crashed and it was still in a stall when flight recordings ended seconds before impact.
At the moment of impact with the sea, Siswosuwarno said the aircraft was "probably" in a horizontal position, with the nose slightly higher than the tail.

http://citraindonesia.com/inilah-det...rasia-qz-8501/

[Lost in Saigon]

Originally Posted by _Phoenix

The plane climbed from 32,000 feet to about 37,000 feet in 30 seconds before entering stall.
The stall warning alarm sounded for four minutes before AirAsia Flight QZ8501 crashed and it was still in a stall when flight recordings ended seconds before impact.
At the moment of impact with the sea, Siswosuwarno said the aircraft was "probably" in a horizontal position, with the nose slightly higher than the tail.

http://citraindonesia.com/inilah-det...rasia-qz-8501/

Here is the Google Translate version of that link:

JAKARTA, CITRAINDONESIA.COM- National Transportation Safety Committee (NTSC) reported the results of the reading of data recorded on the cockpit voice recorder (CVR), one part of the tool on the blackbox, QZ 8501 AirAsia plane that crashed on December 28, 2014 in Karimata Strait, Borneo middle.

From the results of the readings obtained fairly clear picture about the final seconds before the aircraft type Air Bus A3200-200 carrying 162 crew and passengers fell into the sea.

The following data on the moments of the crash AirAsia QZ 8501:

At 6:12 pm:
The pilot asked for permission to air traffic in Jakarta to be able to raise the possibility of AirAsia plane from 32,000 feet QZ8501 38,000 feet. Guides in Jakarta asked AirAsia QZ 8501 stand-by.

At 6:16 pm:
Guides in Jakarta gave permission AirAsia QZ 8501 rose to an altitude of 34,000 feet, but there was no response from the QZ 8501 replies.

At 6:17 pm:
Officers air traffic control lost contact with the plane AirAsia QZ8501 and tried for eight times.

At 6:17:09 pm:
AirAsia plane QZ8501 uphill at a speed of 300 feet in six seconds.

At 6:17:24 pm:
QZ8501 AirAsia aircraft continues to climb. Within six seconds of 1,700 feet altitude increases.

At 6:17:41 pm:
QZ8501 AirAsia plane reaches cruising altitude of 36,300 feet.

At 6:17:54 pm:
AirAsia flight QZ 8501 finally reached the peak height of 37,600 feet before finally experiencing stall.

At an altitude of 32,000 feet, the aircraft position AirAsia QZ8501 down slowly. AirAsia plane QZ8501 slowly down and fell into the sea with an indication of the position of the tail first hit the water.

At 6:20 pm:
Flight data recorder and cockpit voice recorder (black box) AirAsia plane QZ8501 stop recording flight data. (Source: Business)

[Leightman 957]

Argument bruhaha

Triskele: 'At 6:17 a.m. on Dec. 28, three minutes after air traffic control unsuccessfully tried to make contact and asked nearby aircraft to try to locate QZ8501, the A320 turned to the left and it began to climb from its altitude of 32,000 ft (9,750 meters)

Translation: 8501's fate was decided in a very short time.

Gretchenfrage: The question is why do such questions about laws are still unclear after all these years and why do we have so many recipes as to how to deal with such upsets in a bus when ....

IanW: There is no such thing as fault free software.

Mountain Bear, I don't know exactly what you intended by “Experience is not infinite”, but as a statement it points out that all that is known is not enough and never will be enough for the not-yet-encountered. So computers can at best perform no better than what has been discovered so far.

But, Mountain, you also wrote: “There is only one possible exception to that rule and that is...(omniscience)”. That's misleading and untrue, based on your earlier sentence. The total of experience learned to date does not come close to the totality of possiblities. The computer can no further than the programmers' understanding. A320 pilots' misunderstanding about the breadth of that knowledge gap is being demonstrated on this forum.

Confusion is evidenced on these pages about how AC systems are supposed to work, as noted by HeavyMetallist: “it's equally clear that very few of those supposed pilots have a a clue about how their aircraft are designed and engineered”.

This confusion is NOT attached to those who post but lack experience and knowledge (and who also get the sharp end of the stick from people who claim to know). Go ahead and disregard every poster who you think is NOT an A320 pilot, or who is not familiar with A320 systems or their land based or regulatory counterparts, if it makes you feel good. They don't count. What is really pertinent here are the A320 pilots (and their supervisor masters) who disagree. THAT's a problem.

The troubling argument is between A320 cognoscenti who think computer systems at present are adequate, and those who can cite or have experienced examples of why they believe computers have important shortcomings.

[Shadoko]

From the "published" data :
At 23:16:52, ground speed was ~470 kts, that is ~242 meters/second.
At 23:17:29, ground speed was ~311 kts, that is ~160 meters/second.
So, in 37 seconds, the horizontal speed went down by ~82 meters/second.
This is a 82/37 = 2.2 meters/second˛ deceleration. So, in the plane, the apparent vertical was tilted about 13 degrees (arctangent(2.2/9.8) = 12.6). A 13° tilt upward of the plane might be felt as no tilt and constant speed... If the captain was standing up, playing with the breakers, he might be not aware of the trajectory problem until the plane began to fall.
IMHO.