Note: we continue to add latest news to this article, updates are from now on in blue.
March 23, 2015; An A320 from Germanwings, a subsidiary of Lufthansa, has crashed today after contact was lost with the aircraft at 10.47 UTC over French Alps. The aircraft, with 144 passengers and six crew members, was on scheduled flight 4U9525 from Barcelona, Spain to Düsseldorf, Germany. The crash site has been identified north of Dijne-le-Bain in Alpes-de-Provence, French authorities has reported there are no survivors.
The aircraft, an A320, was serial number 147 from 1990, one of the older in the fleet of Germanwings.
Nothing is communicated about a possible reason for the crash, which happened after a steep descent from cruise altitude just after the aircraft reached the French coast east of Marseilles, Figure 2.
The vertical path shows a normal flight until 09:31 (position of aircraft symbol on map) when the aircraft enters a fast descent with over 3,000 feet per minute. The aircraft groundpeed remains constant and then tapers slightly to just under 400kts just before impact with the French Alps north of Dijne-le-Bain. Course is steady during descent at 026°. Place of impact is N44°16’49.80 E6°26’22.20 , enter this in Google maps and press search. Nearest village to W is Le Vernet with a larger village to NW, Seyne, also named Seyne-les-Alpes where most of the rescue teams have put up base.
Figure 3 shows the immediate vicinity of the crash site with closest village Vernet and the road that leads to the closest mountain top, Col de Mariaud. The crash site is in Ravin de Rosé, east of this top.
The French rescue specialists from the military police has searched the crash site and found the flight voice recorder black box, the data recorder is still being searched (first information was the other way around, this has since changed). The aircraft flew directly into the side of a ridge and was virtually pulverized. There are no survivors.
There has been rumors of a distress call from the aircraft, French authorities has confirmed that this is false, as they saw the long descent (8 minutes) and could not get contact with the aircraft they declared an emergency at 10:36 UTC.
The French Bureau d’
The crash site is in a remote area and recovery of debris from the aircraft will take long time especially as the weather, which has been good Tuesday, is now turning worse. The flight voice recorder can give crucial information why the descent was initiated but it is not sure, if events has surprised the crew the recorder gives less information than if the data recorder would have been in the hands of investigators. The search for this recorder will continue Wednesday, crash investigation personnel from the French military police (les Gendarmes) has camped on the crash site to secure the area and to continue the search at day break.
BEA Paris has started the investigation of the voice recorder, it is damaged but readable. BEA has released a picture of the recorder:They will hold a press conference at 15.00 UTC. It remains to be seen how much of the mystery the voice recorder can clear, especially what happened to the crew. It was clear weather in the crash area and they should have seen the approaching alps. If the voice recorder can not bring all pieces needed for clarity the data recorder, which is not yet found, should. The full investigation with final report typically takes a year or more.
The working assumption from Investigators and Lufthansa/Germanwings is that this is an accident and not a terrorist attack. What confounds experts is that no communication or change of transponder code followed on the un-announced and un-allowed descent from FL380, yet the descent was orderly as if programmed into the aircraft autopilot and not the falling out of the sky that a catastrophic even would have caused.
The accident has got a lot of attention, the last airline accident in France was the Concorde crash July 2000. French interior and ecology ministers are on site Tuesday as are Airbus Group CEO Tom Enders and Airbus CEO Fabrice Brégier.
French President Francois Holland, German Prime Minister Angela Merkel and Spanish Prime Minister Mariano Rajoy arrived at Seyne at 13.00 UTC after their helicopter had first circled the crash site. At Seyne they talked at length with rescue personnel before going for a prepared briefing on the rescue operation:
Lufthansa is said to fly victim families to the area tomorrow via Marsielles, the villages Seyne and Vernet has arranged for reception and condolence facilities in cooperation with the authorities in Dijne-Les-Baines.
The weather has been good Tuesday but should have turned worse. Luckily the forecast has not been correct, there is moderate winds and overcast with light drizzle from time to time in the area.
Our thoughts and prayers are with the families and friends of the passengers and crew.
The speed during the descent correspond quite well with MMO/VMO limits of the aircraft, which are Mach .82 and 350 KCAS.
Small correction: it’s Dignes-les-Bains
On the face of it, seems to be a fairly controlled but urgent descent. I wonder if it was precipitated by a cabin decompression event of some sort, although of course that doesn’t explain why control was lost once at lower altitude.
Thanks for the correction. Google Earth didn’t like the misprint. The sad thing on this one was that all on board were likely conscious until the end.
Coincidentally, today a toe-curling report was published where an Airbus flight computer saved a (A330 MRTT) crew / aircraft from crashing by intervening in an extreme situation.
http://www.flightglobal.com/news/articles/a330-flight-control-laws-saved-voyager-inquiry-finds-410491/
Wow. That’s some chilling reading.
One piece of technology (side stick) got them in trouble (grin) and the other helped save them. Can’t happen on a Boeing of course!
Shades of the movie version of “Fate is the Hunter”.
And a big phew that it ended well.
Couple of years ago a GA plane crashed in the Baltic due to the inflatable life rafts ripcord having been mishandled during loading.
During cruise the think inflated unexpectedly in the back and pressed the jokes ( via seatbacks and the pilots) forward. no survivors.
Idiotic behavior by the captain. Perhaps large red stickers are in order, warning the flight crew not to stow personal items, such as camera’s, phones, iPad’s, laptops, coffee mugs, magazines, etc. near a sensitive flight control input device.
Straight line descent to a certain altitude, semi-constant speed, straight line course, no communications. That gives me the shivers..
@roger yes, or fire, or intentional controlled flight into terrain? Leaving all other speculation to the media. My thoughts are with the surviving relatives and friends.
very disturbing. as I understand the story so far, pilot declares IFE at 38kft over Marseille, begins emergency descent but doesn’t change flight path to divert to nearest airport, flies into side of mountain after descending below MSA.
depressurization induced loss of consciousness @38Kft would have occurred (if the Pilots both failed to don O2 masks) within 60 seconds , i.e. before the plane got to 35Kft @ 3Kft/min descent rate. most people are more or less fine below 15Kft, but it would have taken 5 minutes to get there, and even if they did survive that long (as death from hypoxia comes not long after LoC) it would certainly take them some time to come fully awake and aware of the situation at hand, possibly not leaving sufficient time to process the danger and react accordingly.
As far as I could gather there was no IFE declaration coming from the plane. ATC kicked that off after they could not raise the flight on the wireless.
Here is what we know, including speculations of my own.
No emergency was declared by the pilots. The airplane descended more rapidly than a normal descent, but less rapidly than an emergency descent. It looks like it was a computer initiated descent since it was performed step wise, as if the computers had detected an abnormal condition and was bringing the aircraft to a lower altitude in programmed steps. In its semi-fast descent it hit the mountain, so obviously the terrain avoiding system must have been turned off. We can also assume that both pilots were incapacitated at high altitude (38,000′) for one reason or another. This could be related to a pressurization problem. In which case we would have to assume that the pilots did not done their oxygen masks for some reason.
At least one black box has been recovered, so we should have a better understanding of what might have happened in the coming weeks.
“It looks like it was a computer initiated descent since it was performed step wise, ..”
?step wise?
I see the sink rate changing gradually.
The last segment of level data items are imho artifacts from a “lack of data” interpolation algorithm.
The comment that refers to the computer initiated descent in steps was made by Éric de Rivery, spokesperson for the French pilot union called “Syndicat national des pilotes de ligne” (SNPL). The exact sentence in French goes like this: “Un appareil peut de lui-même, sans action volontaire des pilotes, amorcer une descente, s’il détecte une anomalie, pour tenter de la corriger. Ce qui pourrait expliquer la descente par pallier, à priori sans décrochage sec, de l’appareil.”
En savoir plus sur http://www.lesechos.fr/industrie-services/tourisme-transport/0204249863542-crash-a320-quelles-hypotheses-un-pilote-de-ligne-repond-1105014.php?8QKhxLOk3yk2jcdQ.99
Since there is an apparent discrepancy between what Uwe has observed in the graph and what is said in the article, here is how I can best explain this contradiction.
The word pallier may have been misspelled by “Les Echos”. The proper spelling for the preposition ‘par’ should be ‘palier’, not ‘pallier’. The latter is actually a verb. So it is possible that Rivery was misquoted and actually said “pour pallier” instead of “par palier” (misspelled here as ‘pallier’)”. The word ‘par’ means ‘by’ and the word ‘pour’ means ‘for’.
My own impression though is that it was a misspell, not a misquote. Otherwise Rivery, or someone else, would have caught this and asked for a correction. For a misquote has more serious consequences than a simple misspell. The jury is still out.
Just to clarify: I found both image links on heise.de : http://www.heise.de/tp/foren/S-csv-Daten-vom-Flightradar24-vs-Fernsehbilder/forum-294171/msg-26665732/read/
csv data seems to be legit. the graph could have errors, though I don’t think so.
_and the graph shows position over height and not height over time_ .
@Uwe
When I look at the graph on “Les Echos” there is a blue line that represents the altitude and a red line that represents the speed. Clearly, the red line, which represents the speed, shows several steps during the descent phase.
I realize that someone who cannot read French may have difficulty understanding what I said above. So let me add this for the English speaking people.
If Rivery said “pour pallier” it meant ‘to respond’, ‘to compensate’ or ‘to correct’, which all means more or less the same in this context. But if he said (as I think he did) “par palier” (spelled incorrectly as ‘pallier’), it meant ‘in steps’.
Questions that arise, what reasons would there by for the terrain avoiding system not to be working? During an emergency descent, would it not be more appropriate to set the new cruise altitude to say 10.000 feet, just in case you as a pilot would pass out? (see page 17 for other procedures, such as IFE http://www.niceflight.free.fr/private/02_A320FlowPattern.pdf)
That the terrain avoiding system was turned off is hard to explain as, like you suggest, it’s meant to be activated at all time and especially when flying in the Alps. But remains the possibility that it was deliberately turned off. I doubt this very much though, for this accident does not appear to be a deliberate act like in the case of MH370.
I agree with you that there are many potential causes which made the terrain avoidance system inoperable or the airplane unresponsive to its inputs. It is far to early too suggest that it is an deliberate act, but the same counts for dismission of a deliberate act.
I’m glad the data recorder has been found, I assume the French are quick in giving a first insight of the results.
The NTSB would normally be much faster than the BEA. But the fact that the crash occurred in France, and the airplane was certified in France as well, should speed things up a little.
Here are a couple videos of the crash. The comments in French say that the airplane has been completely obliterated (pulvérisé).
http://www.linternaute.com/actualite/faits-divers/1222240-crash-a320-premiere-video-dans-les-alpes/
csv data (position,height : only as a pic unfortunately )
https://twitter.com/flightradar24/status/580332876081811456
and generated plot:
http://www.pic-upload.de/view-26506065/germanwings.png.html
Uwe, look at the light green cuve for speed: it’s in steps all the way.
light green, negative values ~= vspeed ~= vertical speed/sink rate ?
… and I’d call that “fluctuating” not in steps 😉
Uwe, I no longer think that the step levels I was looking for are to be found in this chart. However this speed curve is indicative of an aircraft receiving inputs from a computer as it shows constant fluctuations like if the computers were making adjustments to maintain a smooth descent curve. In other words what is making the smoothness of the altitude curve is actually the roughness of the speed curve.
Now lets get back to the level steps I was alluding to earlier. My current understanding is that after initiating a descent the aircraft apparently levelled off around 10,000′. If this is true it would be a clear indication that the descent might have been initiated in order to bring the passengers to a safer altitude following a cabin depressurization. The aircraft either did this on its own or was commanded by the pilots. But the latter is highly unlikely for the following reasons:
1) No distress call.
2) No communications whatsoever.
3) No response to the ATC repeated calls.
4) No response to a likely warning from the GPWS.
5) No emergency descent profile.
6) No navigational deviations.
excel file containing FR24, FlightAware and Plane Finder data here:
https://www.dropbox.com/s/e2yj8f9x7yjwmsa/4U9525_2015-03-24.xlsx?dl=0
Really frightening, but at the same time really puzling. It is obvious the aircraft has had a very high speed at impact, both from the pictures/videos and also from the Flightradar data. But what strikes me is that there is no sign of fire. All the debries seems white, not black or scorced. Granted, the aircraft probably had not that much fuel on board, but still.
As sad as this accident is it will still be very interesting to learn about the cause in due time.
The mountain sides look a little scorched to me, and first picture do seem to show smoke/steam plums
At that impact speed enough energy is released to vaporise all fuel and provide for rather clean combustion ( think aerosol bomb )
Ok, thanks for the clarification.
I find it a bit odd to comment like that when if you look at the flight path they are maybe 1/4 of the way to Dusseldorf (their destination)
Of course they would have a lot of fuel on board.
Still the collision speed provides enough energy to
atomize/vaporize the contained fuel.
Make an experiment. Throw a bag of kerosene with 150m/s against a wall and see what happens 😉
I’ve done your experiment with water. A ~150 m/s impact into a hard surface does not make droplets as small as you think. Plus, proper dispersion is just as important as droplet size in achieving the equivalence ratio required for a “clean” burn.
It was hard, for me at least, to tell from the pictures whether or not there was any scorching of the ground. The natural color of the surrounding ground seems to have a lot of grey in it.
150m/s, respect. what did you use?
Water probably is not a good equivalent.
For a start the specific enthalpy of vaporization is rather high
in comparison to liquid hydrocarbons.
Anyway, the mountains there are mostly shale ( both as solid base rock and weathered rubble. ) prone to slides. If you crash anything into it it just looks differently broken afterwards 😉
Believe it or not, 150 m/s is not that hard. We have a set of air cannons (think potato gun, but larger) in our lab that can get to those speeds with 6″ diameter light projectiles. However, I used a powder gun for the experiments that I mentioned, and got speeds ranging from 100 m/s to about 500 m/s.
I get what you are saying about the difference in vaporization heat, but I don’t think that is the primary driver. Even if all the fuel were vaporized, which I don’t believe would have happened for various reasons, the vapor would have had to mix with the surrounding air to an unreasonably great extent in order to burn “cleanly”, given the short timescales involved. Dispersion/mixing is the bugaboo here.
The problem for me with the ground in terms of interpreting what happened, is that the color is not just solid grey but heavily streaked with grey. Essentially, the ground looks to me much like what I would expect to see if a more uniformly colored ground was streaked with soot from a significant combustion event.
shouldn’t the debris then also show some soot?
That looks clean. Even in the very first images.
(i.e. no change of it having been washed off by precipitation.)
You’d get some color changes were shale layers were disturbed
( i.e. turning less weathered pieces up and the paths of little avalanches having been triggered from the disturbance )
Yes, I would expect to see soot on the debris as well as the ground. I would also expect to see charring on the debris from the fireball, even more so if the burn was hotter (cleaner) like you suggested. However, slightly charred debris could still look lighter in contrast to the grey background of the ground. It would help if there was a good close-up of the impact crater as opposed to the overall’s of the debris field that we have all seen.
The last commercial loss of hull accident in France was 2008 XL Airways Germany Flight 888T. The a320 was going through testing regimes as part of a handover to Air New Zealand and suffered a stall as a result of insufficient airspeed and water-blocked angle of attack sensors.
Interesting flight path of Air NZ 767 which turned back from flight from Auckland to Cook Islands.
The tracking shows the return flight to wander over half of the North Island before landing normally. Very strange. Local media picked it up but no explanation, which I think we deserve after the wanderings of MH370
flightaware.com/live/flight/ANZ46/history/20150317/2045Z/NZAA/NCRG
Shows normal track now as well as well as the scenic route on return
Routine fuel burn following a turnaround. The pilots detected a brake issue and had to burn excess fuel before landing, so elected for a scenic route. The passengers got an interesting view to go with their delayed flight.
http://www.stuff.co.nz/travel/travel-troubles/67435600/air-new-zealand-flight-takes-unscheduled-detour
I would have thought a fuel dump and direct return would be far quicker and be more useful for passengers than a something google earth can do.
It seems like its good for pilots only, to do some real flying.
Too often the airline operations doesnt mandate quick landing when some technical issue occurs, as who really knows how serious it is. An Airtran flight to Canada from Havana was made to return back to origin rather than the closest airport ( Ft Lauderdale) when they had some issue with fuel and they only just made it !
Pilots can and should overule the guy back at the desk and computer screens for any technical problem
Normand and Bob, where does this idea of “terrain avoiding systems” and them being off or inoperable come from?
Transport category airplanes have (Enhanced) Ground Proximity Warning Systems, to warn of terrain closure. But they are entirely dependent on the pilots taking action. Manually.
In my initial comment I said “In its semi-fast descent it hit the mountain, so obviously the terrain avoiding system must have been turned off.” In other words the fact that the pilots did not respond indicates that the Ground Proximity Warning System was not working. Because if it had been working the pilots would have done something about it. Provided they were still conscious. But nothing indicates they were. The no-response to the GPWS warning may actually be an indication that they in fact WERE unconscious.
My theory is that the aircraft computers may have detected something like a decompression, or a below nominal cabin pressure. Following this the computers initiated a descent to bring the aircraft to a safer altitude. That altitude would be around 10,000′ and that happens to be the altitude the aircraft is reported to have levelled off before it hit the mountain. In this scenario I am assuming the pilots were incapacitated.
The ACARS’s last reported altitude was 10,275′. So I can only assume that this is the point it had finished transmitting its last batch of data. But the aircraft crashed at an altitude of 6,800′. And because of that it is not clear to me what the aircraft did exactly between these two altitudes.
To the best of my knowledge the computers cannot initiate a decent for any reason let alone a de-compression.
Per terrain avoidance, that’s a pilot command, not an envelop protection.
What we need to do is let the authorities get the data and get some real information out.
Other than what’s been listed there is no information to give a clue into cause.
I am taking this information from the “Syndicat national des pilotes de ligne” (SNPL), the French pilot union. Its representative, Éric de Rivery, indicated that the airplane flew different levels. I heard of at least one such level: it descended from 38,000′ and levelled off at 10,000′. Which is indicative of a deliberate manoeuvre to reach an altitude where the passengers can breath.
I must add that de Rivery specifically mentioned that the aircraft computers would have the capability to initiate an automatic descent of this nature, if need be. Wether it did or not is another story. But we should find out in the coming days.
As I have already mentioned in a previous post, above on this thread, her is what de Rivery said in French: “Un appareil peut de lui-même, sans action volontaire des pilotes, amorcer une descente, s’il détecte une anomalie, pour tenter de la corriger. Ce qui pourrait expliquer la descente par pallier, à priori sans décrochage sec, de l’appareil.”
Translated by me it goes like this: An aircraft can initiate a descent on its own, without pilot input, in order to make a correction if an abnormal condition is detected. This could in principal explain why the aircraft descended in level steps in a no abrupt stall condition.
Maybe a better way to put it is all the aircraft computers can do when fully functional is envelope protection .
They are not allowed to take actions other than to protect the airframe. I..e stall avoidance, turn limiting etc.
No envelope protection I know of would have a decent like that.
The only obvious thing is something very strange went wrong, maybe even bizarre twist in the usual sequence of one thing triggers and a second entity compounds it resulting in a crash.
Computers could certainly can do weird things but this would be the weirdest non human one I know of (some A330s have done some pretty strange things as has at least on 777 when the inputs went strange and the reaction was not what was supposed to happen, all software has bugs)
You can Program in a decent, computer cannot initiate it.
I do wonder on the ground track that is quite a way off a direct line to Dusseldorf though no mention of ATC being upset about where they were.
I heard at least one pilot saying that he flew the exact same route the day before and confirmed everything looked normal for this particular flight until it started to descend. However when the aircraft started to descend the air traffic controllers realized that something was wrong because the pilots were not responding to their repeated calls. This is one of the clearest indication we have that both pilots were incapacitated. We can also assume that the GPWS went off and was ignored by the pilots who were possibly unconscious. Other than an uncharacteristic highjack scenario only hypoxia or fire could explain this. But fire is unlikely since the aircraft did not appear to be on fire before the crash. That leaves us with hypoxia as the most “obvious” cause. For me this tragedy has all the characteristics of a computer controlled Helios Flight 522.
OK, I’ll say it…
This looks like intentional controlled flight into terrain.
Just after reaching cruise altitude, one of the pilots could have left the cockpit and then the descent started.
Shades of EgyptAit 990?
Have the pilots’ names been released yet?
That scenario does not match the flight profile very well. For the descent looks more like a well controlled descent: slower that an emergency descent, but faster that a normal descent. And if it’s true that the aircraft levelled off around 10,000′ this would fit well with a recovery altitude after a lost of cabin pressure. All this done automatically by the airplane’s systems while the pilots were (probably) unconscious. The aircraft would have initiated this manoeuvre to bring the passengers to a safer altitude after it detected an abnormal pressure inside the cabin.
what makes you think the A320 can autonomously initiate a controlled descent if it detects cabin depressurization?
from everything I have read (except your assertions) the aircraft will _only_ do what the pilot tells it to do unless the pilot commands it to exceed the flight envelope, in which case it will refuse exceed envelope limits. no mention of autonomous actions other than audible and visual warnings when flight parameters are off norm.
I have already answered this question from someone else in a previous post on this thread. But basically I take this information from what said a representative of the French pilot union.
More the shades of LAM:http://aviation-safety.net/database/record.php?id=20131129-0
Why did the pilots nog change course to the left? Instead of mountain you have big flat fields of sunflowers.
Luckely already FDR is being read out, so much more clarity is around the corner.
My understanding is that it is the CVR that was recovered. If this is true, and provided it is still readable, it will provide lots of informations by the absence of voice communications they will hear, or not hear I should say. If the pilots were unconscious all the way, as I think they were, they won’t hear them talking or working at all. They might hear only two things: the engine noise and the impact with the mountain.
They will also likely hear the repeated warnings from the GPWS and possibly some alerts related to a low cabin pressure or other unusual events.
This A320 had accumulated 58,300h on around 46,700 flights. A lot of cycles on that air frame. This morning I heard that GermanWings pilots do not trust the old planes anymore and asked for replacements that have less hours on them. So far Lufthansa is replacing 11 of them.
Meanwhile LH is deferring new airframe purchases and buying used E-jets instead …
FlightGlobal reports: Germanwings is operating 40 flights with aircraft and crews from Air Berlin, Lufthansa and TUIfly today.
The Cologne-based Lufthansa subsidiary has chartered 11 aircraft from other airlines.
Don’t forget that there is an intermittent strike going against LH pushed by “Vereinigung Cockpit” .
With the deterioration in strike culture in recent years I’d not be surprised if there is a tie in.
“This A320 had accumulated 58,300h on around 46,700 flights. A lot of cycles on that air frame.”
Still well within limits. The original certified limits are 48,000 cycles or 60,000 hours, and there’s a package available (which LH is implementing on some of their A320s as far as I am aware) to extend this to 60,000 cycles or 120,000 hours.
“This morning I heard that GermanWings pilots do not trust the old planes anymore and asked for replacements that have less hours on them.”
From all that I’ve read, this is nonsense, spun out by media from the fact that some crew did indeed refuse to fly. Not because of safety concerns, though, but because they felt they weren’t fit to fly because they had lost friends and colleagues.
“..some crew did indeed refuse to fly. Not because of safety concerns, though, but because they felt they weren’t fit to fly because they had lost friends and colleagues.”
If there are no safety concerns, why did they charter 11 newer planes?
I belive that there is a connection between airasia flight 8501 germanwings flight 4U9525 and the lufthansa incident in november Also there have been incidents accross the airbus families that the aircraft have start descending without any pilot command so i speculate that there is a fault in the airbus fly by wire system and maybe the pitot tubes. If am right millions of people are at risk every day
A polite suggestion: Please do some more in-depth reading before speculating like you do.
Can someone please confirm the correct POI location? The position reported above (N44°16’49.80 E6°26’22.20) does not fit the BEA live report this morning, which put the altitude of the POI at 6175 feet. I have information from another (French) source that puts the location at N44.2908, E6.4728, which is at ~ 6200 feet, consistent with the BEA report this morning. It is an important detail. TNX.