Iced AoA sensors send A321 into deep dive
Thread Starter
Join Date: Jan 2008
Location: Berlin
Age: 56
Posts: 97
Likes: 0
Received 0 Likes
on
0 Posts
Iced AoA sensors send A321 into deep dive
An Airbus A321 operated by Lufthansa went into a deep dive after iced sensors were feeding the aircraft's systems false data. According to a report by German news magazine "Der Spiegel", the plane rapidly descended for minutes, dropping 1000 meters per minute. The crew reportedly only regained control of the aircraft after "switching off onboard computers".
The incident happened to LH1829 enroute from Bilbao to Munich on November 5th, 2014 with 109 peeps on board. Germany's aviation security authority BFU is investigating a "severe incident", the incident will be featured in their latest bulletin due shortly.
AV Herald has a little more detail (apparently, French authorities are a little quicker and more frequent with their bulletins): About 15 minutes after takeoff, LH1829 was climbing through FL310 "when the aircraft on autopilot unexpectedly lowered the nose and entered a descent reaching 4000 fpm rate of descent. The flight crew was able to stop the descent at FL270."
According to AV Herald, the loss of altitude had been caused by two angle of attack sensors having frozen in their positions during climb at an angle. That caused the software to assume a stall. Alpha Protection activated forcing the aircraft to pitch down, which could not be corrected by stick input. The crew disconnected the related Air Data Units and was able to recover the aircraft.
What the German press is labelling a "near crash" scenario apparently isn't an isolated incident: According to "Der Spiegel", Lufthansa records show more than a dozen cases of iced or otherwise blocked sensors. Airbus reportedly offered to replace the sensors in question and updated the software.
Also, the manufacturer issued an emergency airworthiness directive stating that "when Alpha Prot is activated due to blocked AOA probes, the flight control laws order a continuous nose down pitch rate that, in a worst case scenario, cannot be stopped with backward sidestick inputs, even in the full backward position. If the Mach number increases during a nose down order, the AOA value of the Alpha Prot will continue to decrease. As a result, the flight control laws will continue to order a nose down pitch rate, even if the speed is above minimum selectable speed, known as VLS.This condition, if not corrected, could result in loss of control of the aeroplane." Airbus advised airlines to implement procedures for crews to switch off two ADRUs in this case.
The incident happened to LH1829 enroute from Bilbao to Munich on November 5th, 2014 with 109 peeps on board. Germany's aviation security authority BFU is investigating a "severe incident", the incident will be featured in their latest bulletin due shortly.
AV Herald has a little more detail (apparently, French authorities are a little quicker and more frequent with their bulletins): About 15 minutes after takeoff, LH1829 was climbing through FL310 "when the aircraft on autopilot unexpectedly lowered the nose and entered a descent reaching 4000 fpm rate of descent. The flight crew was able to stop the descent at FL270."
According to AV Herald, the loss of altitude had been caused by two angle of attack sensors having frozen in their positions during climb at an angle. That caused the software to assume a stall. Alpha Protection activated forcing the aircraft to pitch down, which could not be corrected by stick input. The crew disconnected the related Air Data Units and was able to recover the aircraft.
What the German press is labelling a "near crash" scenario apparently isn't an isolated incident: According to "Der Spiegel", Lufthansa records show more than a dozen cases of iced or otherwise blocked sensors. Airbus reportedly offered to replace the sensors in question and updated the software.
Also, the manufacturer issued an emergency airworthiness directive stating that "when Alpha Prot is activated due to blocked AOA probes, the flight control laws order a continuous nose down pitch rate that, in a worst case scenario, cannot be stopped with backward sidestick inputs, even in the full backward position. If the Mach number increases during a nose down order, the AOA value of the Alpha Prot will continue to decrease. As a result, the flight control laws will continue to order a nose down pitch rate, even if the speed is above minimum selectable speed, known as VLS.This condition, if not corrected, could result in loss of control of the aeroplane." Airbus advised airlines to implement procedures for crews to switch off two ADRUs in this case.
Already discussed in theses threads/posts:
- http://www.pprune.org/tech-log/50207...ml#post8802005
- http://www.pprune.org/rumours-news/5...ml#post8800516
- http://www.pprune.org/tech-log/50207...ml#post8802005
- http://www.pprune.org/rumours-news/5...ml#post8800516
Airbus are currently working on the fix
How many Bus fvkups due those probes do we need before the probe prob is fixed or the software is fixed to cope (or the pilots are trained to cope with it)?
Oh dear, another example of Airbus 'we always know better than pilots' automation run amuck.
If a Boeing starts to do something unexpected you simply disconnect the autopilot and point it in the right direction, makes you wonder why Airbus couldn't offer that as an option
If a Boeing starts to do something unexpected you simply disconnect the autopilot and point it in the right direction, makes you wonder why Airbus couldn't offer that as an option
Join Date: Dec 2014
Location: Au
Posts: 13
Likes: 0
Received 0 Likes
on
0 Posts
Ruled by machines or Midas!
As a pax such a disclaimer that the plane cannot be controlled seems to me the height of foolishness in trusting automated systems. I wish those designers were on the plane. Automation is an attempt by incapable persons to undermine human expertise. These incapable persons range from beaurocrats, politicians to businessmen. There is no substitute for human expertise period. Take for instance the digital cameras -how ignorant people have become with photography! Similar to above there is not even a manual override setting in many new cameras. At least I could accept that in a camera which is sitting in my hand but I just can't believe the same thing happening in a device where I have to sit inside of it! Capitalism and its consequences.
Join Date: Apr 2014
Location: Miles from where I want to be.
Age: 39
Posts: 220
Likes: 0
Received 0 Likes
on
0 Posts
Let's not get hysterical here. The aircraft didn't perform as expected due to a glitch, the aircraft tried to protect itself from a perceived risk (which was spurious), the pilots took over and landed safely.
It must be emphasised that the aircraft can VERY easily be brought out of this situation by pressing a couple of buttons, the autopilot is lost and auto thrust is lost. The aircraft will go into alternate law with stabilities to assist, and direct law for landing (cessna mode). You're fully in control. The plane goes where you point it to. The engines go fast or slow, as you want them to.
Contrary to what is peddled by the media and the Boeing Brigade on here, pilots do actually know how to fly the plane manually. Yes we fly on automatics but that is because airspace rules often dictate that we do so. That 'children of the magenta line' video is a real curse for the profession. People see it on YouTube and automatically just assume we all sit there with our feet up, sipping coffee and have no sense of responsibility and have zero skills and trust our own lives to a computer.
Automation allows aircraft to be flown safely and have probably saved more people than they harmed. Remember no aircraft is crash proof.
Asiana at SFO and Turkish at AMS remind us that wether it be Boeing or Airbus, wether your thrust levers move or not, wether your aircraft is fly-by-wire or not, active monitoring and a timely takeover when things don't go your way will save the day.
The crew of this incident flight did this and saved the situation, Thomsonfly at BOH did this and again saved the day just in the nick of time.
Get behind the aircraft and it will kill you. It matters not if you eat your crew food from a table, or perched on your lap.
It must be emphasised that the aircraft can VERY easily be brought out of this situation by pressing a couple of buttons, the autopilot is lost and auto thrust is lost. The aircraft will go into alternate law with stabilities to assist, and direct law for landing (cessna mode). You're fully in control. The plane goes where you point it to. The engines go fast or slow, as you want them to.
Contrary to what is peddled by the media and the Boeing Brigade on here, pilots do actually know how to fly the plane manually. Yes we fly on automatics but that is because airspace rules often dictate that we do so. That 'children of the magenta line' video is a real curse for the profession. People see it on YouTube and automatically just assume we all sit there with our feet up, sipping coffee and have no sense of responsibility and have zero skills and trust our own lives to a computer.
Automation allows aircraft to be flown safely and have probably saved more people than they harmed. Remember no aircraft is crash proof.
Asiana at SFO and Turkish at AMS remind us that wether it be Boeing or Airbus, wether your thrust levers move or not, wether your aircraft is fly-by-wire or not, active monitoring and a timely takeover when things don't go your way will save the day.
The crew of this incident flight did this and saved the situation, Thomsonfly at BOH did this and again saved the day just in the nick of time.
Get behind the aircraft and it will kill you. It matters not if you eat your crew food from a table, or perched on your lap.
Join Date: Jun 2007
Location: SAYE
Posts: 281
Likes: 0
Received 0 Likes
on
0 Posts
It must be emphasised that the aircraft can VERY easily be brought out of this situation by pressing a couple of buttons, ...
Although different, Perpignan fatal crash (all crew, no PAX) is another example that comes to mind. There are many others.
Join Date: Apr 2014
Location: Miles from where I want to be.
Age: 39
Posts: 220
Likes: 0
Received 0 Likes
on
0 Posts
There are. But there's many examples of a lack of automation leading to slatial disorientation or a loss of control along the likes of Adam Air or the China Airlines 747 who took a Der Stuka dive off California.
Also. Let's take the Birgenair 757 incident which was similar in that the sensors were blocked and they had unreliable airspeed/air data. I appreciate my synopsis is a very brief summary of a long chain of events.
This was a fairly low altitude incident and there were no high altitude aerodynamic issues facing the crew and plane like AF447. The difference was the A330 autopilot tipped out very early on, the 757 autopilot continued flying on spurious data despite the CPT and FO ASI having over a 100kt difference in airspeed.
The autopilot sensed high speed (and I assume the high speed protections) raised the nose, reduced power, stall induced and the pitch was nose high, the left engine flamed out and the right hand engine at full power flipped the aircraft over.
Both aircraft had similar causal factors. Both had functioning PFDs and engines. Pitch and power with a timely execution of the respective Unreliable Airspeed QRH checks would possibly have prevented a disaster. However I say that in the luxury of my own home and realise that hindsight it wonderful.
It could be argued that automation played a factor in both incidents but it cannot be argued that one manufacturer is better than another. There are no statistics whatsoever to prove one safer over the other.
I believe it is a problem for the industry and not one manufacturer as a whole. My ramblings may appear anti Boeing but I am not. I just tire of the anti Airbus (mostly Americans) rhetoric who go into a frenzy much like feeding Pirahanas whenever there's a chance to have a go. I am just trying to put things into perspective and bring a little balance to the discussion.
The Perpignan incident was the first accident of it's type. I'm quite certain that all airliners have latent faults just waiting for a unique set of circumstances to prevail which will cause another hull loss.
Also. Let's take the Birgenair 757 incident which was similar in that the sensors were blocked and they had unreliable airspeed/air data. I appreciate my synopsis is a very brief summary of a long chain of events.
This was a fairly low altitude incident and there were no high altitude aerodynamic issues facing the crew and plane like AF447. The difference was the A330 autopilot tipped out very early on, the 757 autopilot continued flying on spurious data despite the CPT and FO ASI having over a 100kt difference in airspeed.
The autopilot sensed high speed (and I assume the high speed protections) raised the nose, reduced power, stall induced and the pitch was nose high, the left engine flamed out and the right hand engine at full power flipped the aircraft over.
Both aircraft had similar causal factors. Both had functioning PFDs and engines. Pitch and power with a timely execution of the respective Unreliable Airspeed QRH checks would possibly have prevented a disaster. However I say that in the luxury of my own home and realise that hindsight it wonderful.
It could be argued that automation played a factor in both incidents but it cannot be argued that one manufacturer is better than another. There are no statistics whatsoever to prove one safer over the other.
I believe it is a problem for the industry and not one manufacturer as a whole. My ramblings may appear anti Boeing but I am not. I just tire of the anti Airbus (mostly Americans) rhetoric who go into a frenzy much like feeding Pirahanas whenever there's a chance to have a go. I am just trying to put things into perspective and bring a little balance to the discussion.
The Perpignan incident was the first accident of it's type. I'm quite certain that all airliners have latent faults just waiting for a unique set of circumstances to prevail which will cause another hull loss.
Join Date: Aug 2006
Location: Iceland
Age: 40
Posts: 8
Likes: 0
Received 0 Likes
on
0 Posts
It's a strange coincidence that I just finished watching the Air Crash Investigation episode on the Perpignan crash on YouTube when I swipe over to Google now and see this thread suggested to me... Coincidence or perhaps just Google working their magic. I will be starting my training on the A320/321 next month (first pilot job) and it will be interesting to see whether there is any training directly pertaining to these loss of sensors in flight.
Below the Glidepath - not correcting
Didn't anybody at Airbus think performing some Failure mode, effects, and criticality analysis (FMECA) on the air data probes might be a good idea? Can't believe something this basic got through qualification and certification.
Join Date: May 2005
Location: On a good day - at sea
Posts: 263
Likes: 0
Received 0 Likes
on
0 Posts
LH has experienced more than a dozen of these incidents?
We fly more than 100 A319s, 20s, 21s and 30s in the same climate as LH and I review nearly every every in flight incident we have. In 10+ yrs I've yet to see a single one of these incident reports from our crew.
Why not?
We fly more than 100 A319s, 20s, 21s and 30s in the same climate as LH and I review nearly every every in flight incident we have. In 10+ yrs I've yet to see a single one of these incident reports from our crew.
Why not?
Last edited by nnc0; 21st Mar 2015 at 16:37.
Join Date: Dec 2014
Location: Au
Posts: 13
Likes: 0
Received 0 Likes
on
0 Posts
But the coders are sipping coffee!
A few processes happening in a batch utilizing electronics/mechanics/hydraulics can be of benefit, provided they happen under our watchful eye. I.e., we need to find out the limits to automation.
What is disturbing is reliance on software coding. Ultimately, software codes are human instructions by a proxy, stored in a computer memory. Many a times some of these codes are not executed. Hence the error in them never comes to light. Even if the pilots aren't sipping coffee after pressing a few buttons, the coders certainly are - no one can deny that to err is human. And coders are humans! They have no idea of the relevance, emphasis or importance of individual segments of code. Nor there is a way for them to test it. Blind reliance on software should be abandoned and every automated decision making system should be supervised. As to failure of components - the modern manufacturing scheme of distributed assembly with no quality control but only cost control is to blame!
What is disturbing is reliance on software coding. Ultimately, software codes are human instructions by a proxy, stored in a computer memory. Many a times some of these codes are not executed. Hence the error in them never comes to light. Even if the pilots aren't sipping coffee after pressing a few buttons, the coders certainly are - no one can deny that to err is human. And coders are humans! They have no idea of the relevance, emphasis or importance of individual segments of code. Nor there is a way for them to test it. Blind reliance on software should be abandoned and every automated decision making system should be supervised. As to failure of components - the modern manufacturing scheme of distributed assembly with no quality control but only cost control is to blame!
Join Date: Mar 2015
Location: Exeter
Age: 59
Posts: 161
Likes: 0
Received 0 Likes
on
0 Posts
This is all very true. When using some of my tech (even the cool ones) there will always be a set of button presses or using them in a certain way that will cause the system to crash or to free or to do something random. This is with alpha testing and potentially beta testing with thousands of users.
I don't see why a plane would be any different. There will always be interactions that can cause a problem. Much like when a teacher's fancy power point presentation won't load or crashes and they have to use a black board to do their job. The Airbus is just the same. It is an all singing all dancing PowerPoint presentation with all of the fancy animation, sound effects and other wizardry people like to dazzle their audience with. However if I want an ole' fashion blackboard then I have that option too.
I don't see why a plane would be any different. There will always be interactions that can cause a problem. Much like when a teacher's fancy power point presentation won't load or crashes and they have to use a black board to do their job. The Airbus is just the same. It is an all singing all dancing PowerPoint presentation with all of the fancy animation, sound effects and other wizardry people like to dazzle their audience with. However if I want an ole' fashion blackboard then I have that option too.
Ozmd To be fair, this isn't a software coding error. It appears to be a more fundamental design problem. We have a multiple sensor failure mode that is not detectable except by the resulting behavior of the aircraft. Yes, the crew did an excellent job in responding to the situation. But somewhere in the design process, someone forgot to account for a failure mode affecting multiple systems. And to top it off, there doesn't appear to be any provision in the AoA system design to provide a fail signal immediately to warn the pilot prior to pushing the plane into a dive.
Given the requirements, the coders appear to have gotten their job done right. They have no inputs to tell them the incoming data is bad and now would be a good time to flash a warning rather than cary on with the alpha protection routine. Garbage In, Garbage Out.
Given the requirements, the coders appear to have gotten their job done right. They have no inputs to tell them the incoming data is bad and now would be a good time to flash a warning rather than cary on with the alpha protection routine. Garbage In, Garbage Out.
It doesn't sound like this crew corrected the problem by just 'pushing two buttons'
The report said they had to 'turn off two computers' i've only flown Boeing and Douglas products but It sounds to me like they were pulling circuit breakers to disable these 'protections'
The Airbus attitude of 'we always know better than the pilot' has proven to be a major problem.
There should be just one switch you can operate to disable all 'flight control filtering' and go to direct law, just as is the case on the B777 and B787.
The report said they had to 'turn off two computers' i've only flown Boeing and Douglas products but It sounds to me like they were pulling circuit breakers to disable these 'protections'
The Airbus attitude of 'we always know better than the pilot' has proven to be a major problem.
There should be just one switch you can operate to disable all 'flight control filtering' and go to direct law, just as is the case on the B777 and B787.