← Back to Squawk list
  • 26

Qantas flight terror blamed on computer

Размещено
 
A wild ride mid-flight on a Qantas Airbus that left a third of passengers on board injured as they were tossed out of their seats has been traced to a computer programming glitch, air investigators have found. (www.theage.com.au) Ещё...

blongroad
Новостной маньякМаксималистЭнтузиастНадежный другПочетный членИсследовательФотограф-любительПромоутерСлушательАвтобиограф

Sort type: [Top] [Newest]


MimosaDrive
MimosaDrive 0
In a related article about the payouts to injured passengers, Northrop Grumman was also implicated. They manufactured the Air Data Intertial Reference Unit.
http://www.theage.com.au/travel/travel-incidents/plane-dive-passengers-win-payout-from-airbus-20111220-1p3o7.html
Ответить ↓•Постоянная ссылка•Report
skuttlerats
Jeffrey Babey 0
Ouch! It must have been like being in a popcorn popper for awhile! Thats allot of damage to the ceiling and overhead bins. Again. OUCH!
Ответить ↓•Постоянная ссылка•Report
USAir767
Xander Picot 0
The computer was running Vista.
Ответить ↓•Постоянная ссылка•Report
WALLACE24
WALLACE24 0
Looks like the people with expierence tend to agree that the computers should be there to assist the pilots, but never to take control from them or override them. I will never pilot big iron but from a small plane pilot and pax view that system leaves me cold. I trust the airline pilots or I wouldn't get on in the first place, but I have a hard time putting my fate in the hands of computers. That's for astronauts.
Ответить ↓•Постоянная ссылка•Report
preacher1
preacher1 0
Well, my friend, Boeing, like everyone else these days, puts a lot of stock in automation and that is all based on inputs made for whatever function, mostly while still on the ground. Fly it out, turn on the AP at a predesignated spot and let it fly the plane according to all that input you did(well, not quite that simple but for talking purposes.
BUT, when Mr. Murphy pops in on the flight deck and screws everything up that you have planned, and that AP pops off or you turn it off, it's your Airplane and, to quote an old computer programmer adage, "WHAT YOU SEE IS WHAT YOU GET or WYSIWYG", good bad or plumb ugly. At least you have it with no other influence, and you find out just how good(or bad) a pilot you are.
Ответить ↓•Постоянная ссылка•Вверх•Report
spudtu
spudtu 0
Who is Mr. Murphy?
Ответить ↓•Постоянная ссылка•Вверх•Report
skylab72
skylab72 0
The author of Murphy's Law, "If anything can go wrong, it will."
Murphy-was-an-optimist corollary, "If anything can go wrong it will, and at the worst possible time."
Ответить ↓•Постоянная ссылка•Вверх•Report
WALLACE24
WALLACE24 0
Yea, I could kinda do the same thing in our old Columbia 400 with the GPS'S and the auto pilot. Not exactly an FMS, but the ap can do the work (until it don't work correctly). I realize an airliner has hundreds of computerized functions that may intermingle. Getting that control back from the computer has to be an "absolute" in my mind so you can go WYSIWYG. I kinda think that is why you get paid the big bucks! lol
Ответить ↓•Постоянная ссылка•Вверх•Report
preacher1
preacher1 0
I know one thing; flying pais better than trucking.lol
Ответить ↓•Постоянная ссылка•Вверх•Report
WALLACE24
WALLACE24 0
I know your right!
Ответить ↓•Постоянная ссылка•Вверх•Report
bbabis
bbabis 0
This is strange reading and there is obviously more to the story if someone could enlighten me. I'm not an Airbus guy. The article states that the autopilot was disengaged and five minutes later the aircraft nosedived twice. The crew must have some responsibility for reengaging a faulty system twice. If it was not the autopilot, but some other system such as a stick puller reacting to a perceived stall, then why didn't the same system ultimately put the nose down on AF 447 that would have prevented that accident?
Ответить ↓•Постоянная ссылка•Report
spudtu
spudtu 0
Wow, that's a good question. Coulda used that nose-down on AF447. Painfully ironic.
Ответить ↓•Постоянная ссылка•Вверх•Report
preacher1
preacher1 0
Read that again. It is a little vague but I think after a colose reading, the AP disconnected AFTER the funky readings started. That being said, it don't reall make that much difference. I can't see in this or any previous reports on this or 447, where any notification came to the crew that said THE AP IS OFF,YOU GOT THE PLANE, cause if it goes off, it is in the hands of the pilot. That scares the hell out of me.
Ответить ↓•Постоянная ссылка•Вверх•Report
bbabis
bbabis 0
Wayne, in both accidents the autopilot disengaged on its own due to computer issues prior to any significant control loss and the crews knew it. Not being familiar with the Airbus, my question is; when the autopilot disconnects due to faulty inputs, how much control of the aircraft does a faulty computer still have and how much does a pilot have? In the case of AF 447 it is known that the nose never dropped below the horizon due to a constant nose up command by the pilot. In this Quantas accident, the nose seemed to have dropped despite the pilot's command unless he commanded it down. We're just not getting the full story. I will do my research and come up with some answers, but in the mean time, no Airbus rides.
Ответить ↓•Постоянная ссылка•Вверх•Report
spudtu
spudtu 0
I for one am looking forward to hearing what you find.
Ответить ↓•Постоянная ссылка•Вверх•Report
preacher1
preacher1 0
I will be waiting for that answer as well. The AP did disconnect own it's own. My curiosity is, was there something to the crew that it had disconnected and/or what for. I am like you, kinda puzzled here because if I am reading this thing right, the nose dives came after the AP disconnected, which would indicate that the computer still had some control of something.
Ответить ↓•Постоянная ссылка•Вверх•Report
bbabis
bbabis 0
An experienced Airbus pilot can jump in hear anytime, but this is some of what I found. In the Airbus 330, the model in both accidents, the pilot's side-stick controllers are not connected to the flight controls. Wether on autopilot or manual control the computers ultimately command flight control movement. The pilots get a clear audible and visual warning if the autopilot disconnects and they know they must fly the aircraft. A got'ya can be introduced though when in manual mode. In manual mode the flight control system has two more modes called normal law and alternate law. In normal law the computers have much more control and supposedly protect the aircraft from pilot abuse. It won't let the aircraft stall or pin people to the ceiling if the pilot slams the stick forward. In alternate law the aircraft is much more like a normal aircraft. The pilot can now break the airplane if he/she wants to. I'm not sure how the pilot knows which mode the flight controls are in but clearly, if a pilot thinks he/she has protection but doesn't, bad things can happen. My understanding is that the engines are the same way; no direct link between power lever and engine. The engines could be at a power setting not commanded by the throttles. Maybe an Airbus school would make me feel better but, as it looks right now, I will avoid Airbus if at all possible.
Ответить ↓•Постоянная ссылка•Вверх•Report
spudtu
spudtu 0
Not being a pilot, let alone an AB pilot, I should probably not form a final opinion about these planes. Maybe as you say, if you're typed on one, you can work with the system as well as any. Not convinced of that, but... even as a passenger, I'll avoid them.
Ответить ↓•Постоянная ссылка•Вверх•Report
spudtu
spudtu 0
I see. That spells it out a bit.
Ответить ↓•Постоянная ссылка•Вверх•Report
preacher1
preacher1 0
It's like driving a car. It is our choice as to what we buy or drive, like and dislike but they will all get us from point A to Point B. With Airlines owning both types of planes, they too will get you from point A to point B. If you had a good bid with DAL out of ATL to somewhere and back, and were typed on a Boeing product of some type, all was cool. DAL merges with NWA and inherits their AB fleet and starts integrating them into the system. Then they say, Mr. Pilot, come over here and get typed on this AB. We are going to be running some of these on your run. You say, I don't like them things and they say too bad. Go find you another job. If you like the run and your job, you will learn to fly the plane and you will learn how to fly it well for your own safety, as much as everyone else's but that don't mean you have to like it. Sully has said privately that he wasn't crazy about the Bus but he was a senior Captain with Scare Air and couldn't give that up.
Ответить ↓•Постоянная ссылка•Вверх•Report
spudtu
spudtu 0
I had read about normal law, alternate law and such, but it seemed too inverted to understand. What you report helps clarify things, but leaves me with a sick feeling. Technology is great, but this seems to complicate the pilot's job, not simplify it. I do know that the side-sticks are also not connected to each other, (from the AF447 report) so left and right seats don't know what each other are doing.
Ответить ↓•Постоянная ссылка•Вверх•Report
thunderland2
al fredericks 0
having worked in Q.A engineering. computor codes, cause an effects can NEVER be 100% tested an proven. at best maybe 20%. if one error is found, repaired the overall impact can cause another error somewhere else. on an on. at one time, $1M was offered to any company that could successfully prove even 50% of their software. the problem is also due to compartmentization of effort. not one group knowning the full efforts of the other. LEAVE THE FLYING TO THE PILOTS, NOT THE COMPUTORS. remember, nothing can go wrong, go wrong, go wrong.
Ответить ↓•Постоянная ссылка•Report
WALLACE24
WALLACE24 0
Oh!! Those pesky seat belts are really there for a reason....Now I get it.
Ответить ↓•Постоянная ссылка•Report
spudtu
spudtu 0
Does anybody know, when these computer programmers write the code, is at least their flow chart reviewed by a pilot with many hours in similar aircraft? I've studied aviation, physics, navigation, and I write computer code, but I never had one hour behind the stick. Am I qualified to program a 200-seat Airbus? I hope not. Come to think of it, no matter how many pilots pore over the final code, there is no such thing as software without a bug. Crazy.
Ответить ↓•Постоянная ссылка•Report
preacher1
preacher1 0
I would hope that it is but one of 3 things is happening, based on happenings so far: either reviewing pilot is missing the obvious flaws, Bean Counters are using statistics and making a determination that a change would be too expensive, or, it's not being reviewed at all. Airbus has been burned more than once and people have died because of software problems that could have been headed off with a pilot review but the underlying factor that is causing them problems is trying to automate the process too much!! Automation is fine and can defintely make a pilot's life easier IF HE CAN CONTROL IT RATHER THAN THE OTHER WAY AROUND. A PILOT MUST BE ABLE TO TAKE COMPLETE,MANUAL CONTROL OF THEIR A/C IMMEDIATELY.
Ответить ↓•Постоянная ссылка•Вверх•Report
spudtu
spudtu 0
Wayne, I guess that makes sense. Thanks for your take on that. Seems like they really programmed themselves into a corner, and their ideas about how man and machine should interact is flawed. Maybe it would be way too expensive to program themselves back out of that corner, not to mention it would be an admission of bad design. Very sad state of affairs, considering how much good the technology could do if designed wisely.
Ответить ↓•Постоянная ссылка•Вверх•Report
spudtu
spudtu 0
That must be Airbus' Failsafe Strategy - provide enough room on the flight deck for 5 senior captains, just in case that's what it takes to wrestle the sidesticks back from the computer.
Ответить ↓•Постоянная ссылка•Вверх•Report
preacher1
preacher1 0
LOL!!!!!! man that that is low.LMAO
Ответить ↓•Постоянная ссылка•Вверх•Report
robot40
Joe Daileda 0
Bottom line: Buy Boeing!
Ответить ↓•Постоянная ссылка•Вверх•Report
preacher1
preacher1 0
Sad part of all of it is, I don't know if that is the problem as I am not typed in an AB of any kind. I just know that computer and the automation factor has came up in about every serious Airbus incident that has come along. AF447 will continue to be talked to death, but the engine problem on the Qantas A380 a few months back was compounded by the computer throwing so much crap to the crew. It may have all been legit but had there not been 5 senior captains on that crew, they "wouldn't have had the chance of a snowball in hell," and those are their words, not mine.
Ответить ↓•Постоянная ссылка•Вверх•Report
WALLACE24
WALLACE24 0
These pax now have a clear understanding of negative g's.
Ответить ↓•Постоянная ссылка•Report
preacher1
preacher1 0
lol. I think they got paid pretty good for it too.
Ответить ↓•Постоянная ссылка•Вверх•Report
WALLACE24
WALLACE24 0
Yea, they made more than the pilots on this flight.
Ответить ↓•Постоянная ссылка•Вверх•Report
skuttlerats
Jeffrey Babey 0
Going back on this story leads me to believe that Airbus oughta rename their computer systems to "Hal" Sure seems they're going their way of "2001"
Ответить ↓•Постоянная ссылка•Report
preacher1
preacher1 0
Seems that I read somewhere in that article there were 3 sensors and 1 malfunctioned. Now I don't know if they were redundant or not(apparently not) but if not, there should have been some redundancy built in since those were all critical indications. I don't know if panic here or what but at some point in here there should have been a way for the pilot to turn off all the crap, take control of the plane and say "c'mere heifer". That is assuming that the pilot knew how to fly the plane rather than push the buttons. Strange also that these all happened to the same Airline at the same approximate altitude, off the West Coast of Australia??????????????
Ответить ↓•Постоянная ссылка•Вверх•Report
spudtu
spudtu 0
For that matter, couldn't the computer itself check these other factors when it gets a sudden "drop" in airspeed from the sensors? I suppose that might delay an appropriate response when there really was a drop in airspeed and approaching stall, though.
Ответить ↓•Постоянная ссылка•Вверх•Report
preacher1
preacher1 0
Well, I guess this is the whole part that bothers me. It looks like there would be some type of redundancy of some type for a cross check on these critical readings and/or an alarm(there may be)to alert the crew but so far there has been nothing said about it. There seems to be an alarm for everything else.
Ответить ↓•Постоянная ссылка•Вверх•Report
spudtu
spudtu 0
That was another thing I was wondering about. First on the AF447 and now this Qantas flight. If one or even all airspeed sensors go out of whack, could a pilot check a number of other things, that could give him a clue? If barometric altitude was steady, if GPS confirms constant altitude and course progression, fuel rate constant, maybe that would suggest the sensors were out, and not a true drop in airspeed? I might be using the wrong terms here, I'm not a pilot.
Ответить ↓•Постоянная ссылка•Вверх•Report
acvb
Antonio Carlos Vianna Braga 0
Between Boeing and Airbus, I pick Boeing.
Ответить ↓•Постоянная ссылка•Report
bishops90
Brian Bishop 0
According to the last two paragraphs, we can all be assured that this kind of thing will not happen again (says Airbus), so there.

But.........KEEP YER DANG SEATBELTS ON!!!
Ответить ↓•Постоянная ссылка•Report
preacher1
preacher1 0
That is what they make 'em for.lol
Ответить ↓•Постоянная ссылка•Вверх•Report
tartarus12
Robert Curley 0
Keep it up Airbus. Keep taking the flying out of the hands of the pilots. Airbus won't be happy until one of their planes nose dives from 37k into the ground.
Ответить ↓•Постоянная ссылка•Report
tofutwitch11
tofutwitch11 0
Already happened, kinda?
Ответить ↓•Постоянная ссылка•Вверх•Report
skylab72
skylab72 0
Wow, this one is difficult to visualize! First, 'the computer' and 'the auto pilot' may share and therefore be co-resident on the same hardware. Second, sensor bundles are redundant, but their logical pathways to the digital buss(es) of the computational hardware that uses them, MAY in essence route themselves. Functions, like AP or yaw flight control, are mobile and may 'run' in different 'places' at different times. I have never seen the hardware or software block diagrams for an Airbus, but modern digital control systems often (dare I say usually) digitize analog stimuli early, then manage it in the digital domain. In non-geek, just don't try to analyze these failure modes as a hardware issue. The failure may be analog, the sensors themselves, but the critical failure mode is exactly what Wayne has been saying all along. The DEFAULT controller for the aircraft MUST be the pilot. There must be a simplified DIRECT control path with stand alone proven assist algorithms available for the pilot, instantly on demand! Totally bug free software is a mathematical improbability approaching one, it just doesn't happen in systems as complex as flight control systems. I am sure enough that arrogent software engineering (think Micro$oft, Xander), has blossomed within AB, that I reroute to avoid AB equipment, period.
Ответить ↓•Постоянная ссылка•Report
preacher1
preacher1 0
'nuff said.lol
Ответить ↓•Постоянная ссылка•Вверх•Report
jmilleratp
jmilleratp 0
Airbus is getting scarier all the time. You really got to wonder if their whole let-the-computer-fly paradigm has ever been scrutinized to any extent. I have an increasing feeling that it hasn't, and we have seen what happens as a result.
Ответить ↓•Постоянная ссылка•Report
preacher1
preacher1 0
John: if that is a recent picture, you look young enough that you may have to fly one of those things one of these days and if so, you have my sympathy. That being said, they fly every day and take off/land safely. My career was spent on Boeing and now on the fill in, I get a diet of Boeing and various RJ's and I am not going to be typed in an AB. If it comes to that, I'll just totally quit. The biggest problem that I see and what scares me the most on the AB is the ability of that Computer to lock the Pilot out of the loop if he tries to exceed the envelope, and that tad of excess may be just what you need to pull your tail out of a crack. Been there, done that. Boeing has automation and fly by wire on its newer stuff as well but it is there for the pilot to use if he chooses to, he doesn't have to.
I THINK THE "COMPUTER FLY THE PLNE THEORY" HAS BEEN SCRUTINIZED BUT PROBABLY NOT BY ANYONE THAT HAS EVER SPENT ANY TIME ON A FLIGHT DECK. Those that haven't have no idea what happens up there from time to time. If they have spent all their time as PAX they have never experienced an "Oh S$%^" moment in the left seat.
Ответить ↓•Постоянная ссылка•Вверх•Report

Вход

Забыли пароль?
Нет учетной записи? Зарегистрируйтесь сейчас (бесплатно) и получите доступ к конфигурируемым функциям, уведомлениям о статусе рейсов и другим возможностям!
Стать частью FlightAware
Вы знаете, что реклама помогает FlightAware в отслеживании рейсов?
Вы можете внести свой вклад в бесплатную работу FlightAware, разрешив показ рекламы на FlightAware.com. Мы следим за тем, чтобы наша реклама была полезна и не мешала работе с сайтом. Вы можете быстро включить рекламу на FlightAware или приобрести привилегированное членство.
Отменить