OT. - Great Article re. Air France 447 crash

[Mega]

Re: OT. - Great Article re. Air France 447 crash

The Flight engineer was SO smart as to cut the power to 2 engines!!!!!!.............Oh he didn't tell the capt either.

[BiscayneSunrise]

Re: OT. - Great Article re. Air France 447 crash

Hey Mega, Yes a few high profile accidents by the French. Not enough to condemn entire group, I think. As for the Concorde crash, the flight engineer cut power to the number 2 engine per the captain's order. Standard procedure with a fire warning bell ringing.

If I were to condemn an entire group, I'd look down my nose not at the French pilots but at the Airbus. But I'm biased. As they say "If it ain't a Boeing, I ain't going"

[lakedaemonian]

Re: OT. - Great Article re. Air France 447 crash

Hey Mega, Yes a few high profile accidents by the French. Not enough to condemn entire group, I think. As for the Concorde crash, the flight engineer cut power to the number 2 engine per the captain's order. Standard procedure with a fire warning bell ringing.

If I were to condemn an entire group, I'd look down my nose not at the French pilots but at the Airbus. But I'm biased. As they say "If it ain't a Boeing, I ain't going"

Thanks for the subject matter expertise.....

Unfortunately, I've been flying exclusively on Airbus products lately.....definitely prefer the 777ER...and look forward to the new city pairs that the 787 will open up(hopefully)....which could see my 4 day round trip from NZ to NY cut a bit with a deeper reach to Houston/Chicago from the current LAX/SFE.

I also recall a lesson learned from KAL801 that may be relevant.

Rigid adherence to chain of command and authority...even when a failure in command was clearly detected by the rest of the flight crew...the flight crew continued to rigidly adhere(to their death) to a clearly failing captain.

An atmosphere that prevents the ability to manage up or reasonably question command without persecution eliminates the ability to fault check chain of command.

[Mega]

Re: OT. - Great Article re. Air France 447 crash

[BiscayneSunrise]

Re: OT. - Great Article re. Air France 447 crash

OK Mike, That was funny! You got me on that one.

Lake, don't stress too much about the Airbus. They are fine airplanes. The A320, for instance, compared to the 737 is a much better experience for the customer. The 737 is a 50 year old design, really and Boeing has has pushed the 737 stretch about as far as it can go. It's just that the Airbus design philosophy about interface between pilot and computer is different than what most pilots are used to. It takes a while to re-learn and get comfortable with it.

As for strict command hierarchies, about 30 years ago, the airline industry recognized how stifled communication with authoritarian leaders leads to accidents. The first alarm bell was the crash at Tenerife in 1974 where a KLM and Pan AM 747 collided on the runway, killing 400+. Then in 1980? a United DC-8 ran out of gas while holding for Portland, Oregon. In both instances, the junior crew members knew something was wrong but the captain was too fixated to listen. The Portland crash was the catalyst for human factors training for crews, now know as CRM, Crew Resource Management. Since institution, there hasn't been a crew caused fatality at United in over 30 years.

Although, I think with AF 447, they had the opposite problem. There was no one clearly in charge which is just as bad as being too authoritarian. Just because you flatten the hierarchy doesn't mean you can have a commune. There still need to be a clear leader with authority to make the final decision.

As an interesting aside, front line health care providers and surgical teams are getting the same kind of human factors and leadership training. Medicine can be very hierarchical causing errors when junior members aren't empowered to speak up. Health care is adapting a flatter hierarchy along with checklists and more standardized procedures causing dramatic reductions in error along with much better patient outcomes.

[skidder]

Re: OT. - Great Article re. Air France 447 crash

A little background on me: I've been in aviation all my adult life and worked closely with jet transports (mostly Boeing) and their crews.

The main takeaway from this accident is that many modern airline pilots, especially in Europe and Asia, have been trained specifically and solely to fly big jet transports.

Way back in the olden days, student pilots started out in basic machines an learned basic airmanship and aerobatics with just the wind in their face and only the most rudimentary of instruments. Pilots developed a certain amount of "seat of the pants" feel for they airplanes. Civilian or military, it didn't matter, the syllabus was pretty similar. It was only after demonstrating mastery of airmanship did pilots advance to instrument training ,and even later the use of autopilots and autoflight systems. This is still pretty much the model in the US but elsewhere win the world where light aircraft flying is too expensive they have been moving more to simulator based training. For that matter, now, even the US Air Force is training the UAV pilots now with only very minimal training in a real airplane,

About 10 years ago, the training model for airline pilots outside of North America changed. Youngsters were hired by some airlines at age 19 or so and trained from day 1 with the understanding there were to be flying big jets, just straight and level. There was only a minimal amount of training in real airplanes and even then very little emphasis on basic airmanship. The emphasis was on operating all the computer systems that made the airplane more efficient. After a very brief portion in a real airplane much of the balance of the training was accomplished in a simulator. Some of these students never developed that "seat of the pants" intuition.

The design philosophy of the autoflight system at Airbus is notorious for designing the pilot out of the system. Their reasoning goes that if we place enough safeguards in place, the pilot only has to give general direction to the computer; everything will be foolproof. Airbus places the pilot secondary. in Boeings their autoflight systems are seen much more as a tool for the pilot to use, not a replacement for intuition and skill.

So in AF 447, you had at least the junior copilot and maybe both who had trained mostly in simulators and had accrued thousands of hours of flight time only watching the autopilot fly. A couple of contributing factors: 1)They were probably distracted and flew into the thunderstorm rather than deviate. 2) In the storm the anti ice heat of the sped sensors failed. Virtually unheard of in modern airplanes. Then when the pilots were faced with a one in a billion failure, in a high stress situation, they didn't have the raw airmanship skills to figure it out and react properly.

Training departments and flight schools around the world are using this accident as a clarion call. I'm not certain what changes are in the works exactly but I expect a return to basic airmanship as a key part of the syllabus.

I'll jump in with a couple of comments (6 yrs pilot navy, commercial career as 747 & 757 captain for U.S. major carrier)
I agree with the above commentary in general but thought I'd add some flavor with regard to the idea they were negligent in flying into a thunderstorm.
They obviously saw something on the radar that prompted them to call back to the FA's to take their seats, that's my assumption though. But, the problem with assessing storms on radar is that it takes some experience to interpret what you are seeing and also that not every storm paints. Unfortunately wx avoidance is still a mix of art and science, at least with the radars I've used. You can't fly across oceans and avoid every cloud, that pitot tube failure could have happened in any cloud with the right icing conditions. I've wandered into dissipating thunderstorms (which is what they do after the sun goes away for the most part) that never painted on the radar. I'd hazard a guess that just about everyone whos flown internationally for any length of time probably has. A weather briefing that is several hours old or even one hour old is mostly used as a heads up for turbulence briefings, etc. The main work of avoidance is done using the radar and the radar is on pretty much the whole flight when flying at night. So, with all that said, it's not clear to me that they intentionally, or thru negligence, flew into a thunderstorm. Their experience with St elmos fire and the smell you get from warm moist air going through the air exchangers is pretty common when you do get into that type of storm. The up and down drafts are indicative of convective activity so I'm sure at a minimum it was a dissipating storm but not necessarily one that was painting (or it may have been painting green but not red, so not considered dangerous). Also, if it was just a cell or two the crew would normally steer around them as they go through the area so I'm conflicted as to what they were actually doing to avoid these areas or what they saw because most crews will be discussing the avoidance plan as they approach the area of storms and it doesn't seem like that was being done, unless I missed that part.

The usual protocol when you get into one of these old storms (assuming you couldn't avoid it in the first place) is to disconnect the auto throttle as the air temp is usually gyrating up and down which causes the mach/air speed indicator to gyrate around and the auto throttle then tries to compensate with large power corrections either up or back. Anyway, auto throttle goes off, speed is reduced to optimum turbulence penetration speed for that particular type airframe. If the auto pilot is left on then usually disengage the altitude hold and use the auto pilot manual control to keep the altitude assigned as best as possible given the associated strong up/down drafting. It normally doesn't last but a few minutes. With the auto pilot kicking off because of the pitot tube freezing and the loss of airspeed indicator, you are pretty much just leaving the power set roughly were it was prior to entering the storm, maintaining an attitude (making small inputs) that roughly keeps you around your altitude if in rough wx and steering away from any storm returns being painted on the radar. In no instance should a pilot accept a 7,000fpm rate of climb without pushing the nose over to attempt altitude control, especially with a bad airspeed indicator, doesn't make any sense. They were already too heavy to climb to a higher altitude and anyone whos executed a heavy weight altitude change knows it is done at a rate of climb closer to 500fpm when done at cruise speed, so they must have known they were trading airspeed for altitude. It appears that neither pilot mentally acknowledged that they were partial panel and needed to take affirmative action to fly the airplane using the functioning indicators.

In short (IMO), they may have had their hands on the controls but they weren't flying the plane, they were just watching what it was doing right up until it went into a stall.

As an aside, I've never flown an airbus but I have jumpseated on them quite a bit. I just don't like the way the power levers don't always move with power adjustments, etc. Just seems that much harder to assess what the airplane is doing. Give me a boeing anyday.

[c1ue]

Re: OT. - Great Article re. Air France 447 crash

About 10 years ago, the training model for airline pilots outside of North America changed. Youngsters were hired by some airlines at age 19 or so and trained from day 1 with the understanding there were to be flying big jets, just straight and level.

I'm sure this is true, but I also wonder just how much of this philosophy is due to a focus on selling Airbus products?

Unlike the United States and the Soviet Union, most other nations simply don't have a lot of pilots - private or military.

The sheer lack of private airplanes and military aviation leads to a dearth of opportunities to gain experience. One reason why so many nations have national airlines is in fact to grow a pilot force - an airline is one way to get 'flight time' which otherwise is both expensive and economically non-productive if in a purely military format.

[Chris Coles]

Re: OT. - Great Article re. Air France 447 crash

In my case, I am a Lasham Gliding Society www.lasham.org.uk Blue Card, as they say, with everything including Field landings, aerobatics and navigation. Also FAI Diamond Distance and Goal. Not got my height diamond.... yet!

In gliding we have a very heavy emphasis on slow speed manoeuvring including incipient and full spins. Unless you have not only been shown the wide variety of entry cases and then practised them, you cannot know the full potential. So I agree with the tone of this debate; all pilots should be taught basic airman ship before moving onto any commercial air transport.

[GRG55]

Re: OT. - Great Article re. Air France 447 crash

...As for strict command hierarchies, about 30 years ago, the airline industry recognized how stifled communication with authoritarian leaders leads to accidents. The first alarm bell was the crash at Tenerife in 1974 where a KLM and Pan AM 747 collided on the runway, killing 400+. Then in 1980? a United DC-8 ran out of gas while holding for Portland, Oregon. In both instances, the junior crew members knew something was wrong but the captain was too fixated to listen. The Portland crash was the catalyst for human factors training for crews, now know as CRM, Crew Resource Management. Since institution, there hasn't been a crew caused fatality at United in over 30 years.

Although, I think with AF 447, they had the opposite problem. There was no one clearly in charge which is just as bad as being too authoritarian. Just because you flatten the hierarchy doesn't mean you can have a commune. There still need to be a clear leader with authority to make the final decision.

As an interesting aside, front line health care providers and surgical teams are getting the same kind of human factors and leadership training. Medicine can be very hierarchical causing errors when junior members aren't empowered to speak up. Health care is adapting a flatter hierarchy along with checklists and more standardized procedures causing dramatic reductions in error along with much better patient outcomes.

Biscayne: For many years now I've been under the impression that within the FAA and the airlines the origins of CRM, although it probably wasn't called that at the time, date back to the 1972 Eastern Airlines 401 accident in Florida...and the effort is intended to try to address both conditions you describe above?

What amazes me is that there are still periodic incidents where a lack of adequate CRM and systemic organizational issues, as you describe in another post on this thread, remains a significant contributing factor. The August 2000 Gulf Air accident is yet another classic example of a flight crew flying a perfectly good airplane into the ground [or in that case the water].

[GRG55]

Re: OT. - Great Article re. Air France 447 crash

A little background on me: I've been in aviation all my adult life and worked closely with jet transports (mostly Boeing) and their crews.

The main takeaway from this accident is that many modern airline pilots, especially in Europe and Asia, have been trained specifically and solely to fly big jet transports.

Way back in the olden days, student pilots started out in basic machines an learned basic airmanship and aerobatics with just the wind in their face and only the most rudimentary of instruments. Pilots developed a certain amount of "seat of the pants" feel for they airplanes. Civilian or military, it didn't matter, the syllabus was pretty similar. It was only after demonstrating mastery of airmanship did pilots advance to instrument training ,and even later the use of autopilots and autoflight systems. This is still pretty much the model in the US but elsewhere win the world where light aircraft flying is too expensive they have been moving more to simulator based training. For that matter, now, even the US Air Force is training the UAV pilots now with only very minimal training in a real airplane,

About 10 years ago, the training model for airline pilots outside of North America changed. Youngsters were hired by some airlines at age 19 or so and trained from day 1 with the understanding there were to be flying big jets, just straight and level. There was only a minimal amount of training in real airplanes and even then very little emphasis on basic airmanship. The emphasis was on operating all the computer systems that made the airplane more efficient. After a very brief portion in a real airplane much of the balance of the training was accomplished in a simulator. Some of these students never developed that "seat of the pants" intuition.

The design philosophy of the autoflight system at Airbus is notorious for designing the pilot out of the system. Their reasoning goes that if we place enough safeguards in place, the pilot only has to give general direction to the computer; everything will be foolproof. Airbus places the pilot secondary. in Boeings their autoflight systems are seen much more as a tool for the pilot to use, not a replacement for intuition and skill.

So in AF 447, you had at least the junior copilot and maybe both who had trained mostly in simulators and had accrued thousands of hours of flight time only watching the autopilot fly. A couple of contributing factors: 1)They were probably distracted and flew into the thunderstorm rather than deviate. 2) In the storm the anti ice heat of the sped sensors failed. Virtually unheard of in modern airplanes. Then when the pilots were faced with a one in a billion failure, in a high stress situation, they didn't have the raw airmanship skills to figure it out and react properly.

Training departments and flight schools around the world are using this accident as a clarion call. I'm not certain what changes are in the works exactly but I expect a return to basic airmanship as a key part of the syllabus.

Difficult to argue with your observation that the efforts by commercial aircraft designers to limit the potential for human error to inadvertently put the aircraft outside the design flight envelope comes with some potentially fatal side effects when those systems fail, or have what might be considered a designed in fault. The rudder-input induced separation of the vertical stabilizer that caused the American Airlines Flight 587 A300 accident in Queens, NY, might be considered an example of the latter.

That is only exacerbated by the natural human tendency to disbelieve something that "can't happen", even in the face of overwhelming evidence that it has. With all the system redundancies built into a modern airplane it seems the potential for pilots to initially accept and then deal with the consequences of a failure of one of the systems [like the pitot-static system in this instance] is perhaps reduced? The 2009 Turkish Airlines B737-800 accident at Schiphol where the radio-altimeters and auto-throttles contributed to stalling the airplane on short final would seem another example of a crew failing to "fly the airplane" after a system failure, and is consistent with what you are describing as basic piloting skills in your posts on this thread.

The preliminary findings in the Boeing 737 accident, released at a news conference in The Hague, prompted the manufacturer to issue an unusual world-wide alert covering procedures that already should be second nature to aviators: Boeing Co. reminded pilots "to carefully monitor primary flight instruments during critical phases of flight" such as takeoffs and landings.

I have never flown a commercial airliner, but in defence of the Air France pilots is it even possible to detect the onset of an aerodynamic stall in a swept wing jet in cruise configuration at high altitude without the aid of instruments [e.g. from airframe buffeting, etc.]? Like any piece of modern engineering the wing of a commercial airliner is a compromise between efficient cruise flight in the transonic region and the low speed flight needed for safe take-offs and landings. To achieve the latter the typical airliner wing is equipped with a mechanical nightmare of moving parts including flaps and leading edge devices. As I recall from my engineering student days, a swept wing can have extreme non-linear pitch moment curves at low speeds, but I don't remember if that is also a factor at the high angles of attack in a stall.

Edit added: Back in the 1960s when I was a kid fascinated with airplanes [and the space program], but too young to do anything about it except build models, I spent one summer in the Fraser Valley east of Vancouver working on my grandfather's farm, which was just a couple of miles on the backside of the Abbotsford airfield. Among other things, Vancouver based Canadian Pacific Airlines used Abbotsford to train pilots and it was a common site to see a DC-8 doing circuits with stop-and-go and touch-and-go landings. Can you imagine anybody doing that today Biscayne?

That was a formative summer for an aviation-mad kid. During the annual airshow in early August, I hopped the back chainlink fence to gain access [airfield security wasn't what it is today, that's for certain]. It was the first time I got to watch Bob Hoover in his twin piston engine Shrike Commander, and Dr. Art Scholl perform aerobatics in his Super Chipmunk with Gershwin's Rhapsody in Blue playing over the tinny loudspeaker system. It was the height of the Vietnam war and the US Navy Blue Angels were flying McDonnell Douglas F-4 Phantoms [about the only airplane I can think of that would be worse than an F-4 for close formation flying might be the Lockheed F-104 :-)] Thunderingly loud and heady stuff for me. For anyone who wants to see some truly exceptional "basic airmanship" here's a link to an old promotional film of Hoover performing in the Shrike. The both-engines-out loop, roll and alternating wheel landing at the end should make any pilot jealous. Hoover's Shrike is now in the Udvar-Hazy Center of the US National Air and Space Museum in Virginia. You'll find the glimpses of the instrument panel rather quaint in this day of all-glass cockpits Biscayne :-)

http://www.youtube.com/watch?v=g7R7j...eature=related

[GRG55]

Re: OT. - Great Article re. Air France 447 crash

A little background on me: I've been in aviation all my adult life and worked closely with jet transports (mostly Boeing) and their crews.

The main takeaway from this accident is that many modern airline pilots, especially in Europe and Asia, have been trained specifically and solely to fly big jet transports.

Way back in the olden days, student pilots started out in basic machines an learned basic airmanship and aerobatics with just the wind in their face and only the most rudimentary of instruments. Pilots developed a certain amount of "seat of the pants" feel for they airplanes. Civilian or military, it didn't matter, the syllabus was pretty similar. It was only after demonstrating mastery of airmanship did pilots advance to instrument training ,and even later the use of autopilots and autoflight systems. This is still pretty much the model in the US but elsewhere win the world where light aircraft flying is too expensive they have been moving more to simulator based training. For that matter, now, even the US Air Force is training the UAV pilots now with only very minimal training in a real airplane,

About 10 years ago, the training model for airline pilots outside of North America changed. Youngsters were hired by some airlines at age 19 or so and trained from day 1 with the understanding there were to be flying big jets, just straight and level. There was only a minimal amount of training in real airplanes and even then very little emphasis on basic airmanship. The emphasis was on operating all the computer systems that made the airplane more efficient. After a very brief portion in a real airplane much of the balance of the training was accomplished in a simulator. Some of these students never developed that "seat of the pants" intuition...

...

In the June 2010 issue of Flying magazine, airline pilot and columnist Les Abend's contribution was titled "Systems Manager". It described his transition from the B767 to the B777, and feeling almost insulted at thinking he was becoming the operator of a "650,000 pound computer". He describes the benefits of the massive amounts of redundancy in the 777, the safety systems such as the repeated automated re-lights in the event of an in-flight engine failure, and the functioning of the EICAS and FMS systems [I found it a bit ironic when he writes about the FMS automatically tuning the frequencies in our now "ancient" VOR system :-) ]. At one point in the article he admits "...it was becoming apparent that I would never know every aspect of the 777. I would have to be satisfied with only acquiring enough knowledge to keep 250 passengers safe. The rest I would learn over time. I'm not even sure the Boeing engineers know exactly what they created..." which seems to fit with some of what you are telling us.

BTW, the same issue has an Aftermath article by Peter Garrison about Colgan 3407. In that instance there was no failure; the autopilot disconnected when the stickshaker warned of the impending stall, as it was designed to do. It seems the pilot was surprised and incapable of hand flying the airplane, and fought the stick pusher all the way to the ground. I remain amazed that such a basic error can be made by a professional pilot. Demonstrating recovery from not only stalls, but also a fully developed spin recovery is a mandatory requirement in the Canadian private pilot flight test. After I soloed I used to take the schools C150 out and do spins for fun [I was much younger then :-) ]. About a decade ago I took an unusual attitudes course in a piston Beech T-34 at a school that used to operate out of DW Hooks in Houston, Tx. I get the impression that there's a lot of professional pilots out there that have never experienced that sort of thing???

[Chris Coles]

Re: OT. - Great Article re. Air France 447 crash

Now if you want to have some real fun, try a 19 metre Janus two seater glider. This is a late seventies, early eighties glider that has a relatively small tail fin combined with relatively heavy wings and if you wish, take it to a safe height, say above 3000' and try a vigorous entry into a side slip. It will go completely sideways and stop flying, stalling away towards whatever direction you have turned.

I will always remember an instructor that showed me, in quick sequence, seven different entries into an incipient spin from slow airspeed; one after the other. The flight envelope of an aircraft must be fully explored before any pilot is let loose with paying passengers. However, once carefully trained, the response becomes second nature..... and fun too.

[BiscayneSunrise]

Re: OT. - Great Article re. Air France 447 crash

Here is a link to the French Bureau of Inquiry final report on Air France 447. Go to "Final Report". Scroll to page 197 for the summary and findings.

http://www.bea.aero/en/enquetes/flight.af.447/rapport.final.en.php

Here is my summary of what happened:

The Airbus 330 had a chronic history of unreliable airspeed indications using the pitot system designed by the Thales company with 13 prior events at different airlines (BEA, 2012). Air France recognized this problem and began discussions in November, 2008 with Thales and Airbus on a fix or a replacement using a system designed by Goodrich. Airbus asked Air France to consider waiting while Thales conducted tests and a redesign. After a few months AF insisted on a replacement and began refitting their fleet with the Goodrich pitot systems just six days prior to the accident.

The accident airplane was still using the old Thales system and entered an area of thunderstorms with the two copilots flying while the captain was in the rest bunk. Immediately upon entering this area, the airspeed became unreliable causing the autopilot, auto thrust and flight directors to fail. The loss of reliable airspeed data caused the flight control system to reconfigure into alternate law, which did not provide normal stall protections. The pilot flying (PF) in the right seat responded by pitching up sharply causing the airplane to enter a stall. The pilot not flying (PNF) briefly overrode the incorrect inputs of the PF using the priority switch on his control stick but the PF, in turn, overrode the PNF inputs using his own priority switch. The Airbus is designed so that the second pilot to press the override switch gets priority. The PF held pitch up input for the duration of the event with angle of pitch as high as 15 degrees and an angle of attack as high as 40 degrees. From the time of the airspeed failure to impact was just four minutes and 23 seconds.

Human Factor related causes were listed as (BEA, 2012):

-Pitot icing in the Thales system a known but misunderstood problem
-Company procedures assume basic airmanship and that pilots can rapidly diagnose and manage a deteriorating condition
-Company training also assumes that pilots are always in control of the aircraft and a rapid diagnosis of problems will allow the correct use of a checklist
-A double failure showed the limits of the current safety model
-The failures caused a complete surprise to the pilots
-Lack of any training in high altitude airplane maneuvering
-Startle effect caused pilots to mismanage the situation and created a highly emotional state
-Failure to identify the stall was attributed to a low exposure in training of this type of problem
-Confusing the stall buffet with a high speed buffet
-Crew didn’t recognize or understand the implications of the aircraft reconfiguring into alternate law
-Over concentration on spurious visual indications with aural warnings being discounted
-At the beginning of the event, transient stall warnings were considered spurious
-Flight director indications were intermittent and inappropriate misleading the crew
-Lack of a clear display showing invalid airspeed indications
-Flight director intermittent indications led crew to believe their actions were appropriate when they were not
-Crew not recognizing the stall and lack of immediate response to recover
-Excess control inputs by the PF caused further destabilization of the aircraft.
-The crew became progressively distracted and never realized the solution was a simple one.
-Complete loss of cognitive awareness causing increasingly weakening task sharing
-Ergonomics of a poor design

[shiny!]

Re: OT. - Great Article re. Air France 447 crash

That blackbox recording is so depressing. I've heard commercial pilots refer to the Airbus as the "French Lawn Dart". I seldom fly, but when making a reservation I ask what type of plane will be used; if it's an Airbus I take a different flight. Perhaps they've fixed the problems and now train their pilots better? I don't know...

[lektrode]

Re: OT. - Great Article re. Air France 447 crash

the "French Lawn Dart". I seldom fly, but when making a reservation I ask what type of plane will be used....

and THEN you have the CRJ's.... esp the ones mesa flys.... 'landing' isnt quite the word i use to describe touchdown..
altho they are cheaper than my other option, i cant keep "if it isnt a boeing, i aint going" out of my head....
but i do like the screens on the back of the seats that the airbus' have