A350 certified; we provide a comprehensive program review

Airbus’ most recent all-new aircraft, the A350 has achieved civil airline airworthiness certification from EASA today September 30, FAA certification will follow. It marks an end to an eight year program to develop an all-new airliner in the 250-350 passenger segment. It also creates a point where a review of this last (for quite a while) big aircraft program is called for.

Airbus photo of A350 test aircraft in formation, celebrating the certification.

Airbus photo of A350 test aircraft in formation, celebrating the certification.

Below we go through all the aspects of the A350, not only program and technical aspects but also organizational, economical and market communication aspects. In all those dimensions it was the big step forward.

The first incarnation of the A350 was conceived as a response to Boeings 787 during summer 2004, initially as a revamped A330 (very much like today’s A330neo) and then finally as a re-winged, re-engined A350, very similar in its changes to today’s 777X.

Airbus slide from 2005 showing the main data for the then A350

Airbus slide from 2005 showing the main data for the then A350.

When the program started in its final iteration at the Farnborough Air Show in 2006, 53% of the airframe and wings were composites. The concept had been stretched from an eight abreast design to nine abreast, then called A350 XWB for extra wide body. Final industrial launch was 1 December 2006.

A350 XWB design
The XWB variant did away with the A330 fuselage cross section and introduced a carbon fiber re-enforced plastic (CFRP) fuselage which now had a 0,35m or 14 inches wider cabin to enable nine abreast seating with 18 inch seats. Much has been made about the A350 fuselage now being made of large panels, which has four longitudinal joins for the panels to form a barrel section. The competing Boeing 787 was made with shorter barrels which were only joined circumferentially. This is probably the difference between the design that has the least operational and technical influence between the aircraft. They are both CFRP based fuselages which enable higher cabin pressure and humidity for passenger comfort and reduced maintenance requirements as CFRP is virtually fatigue and corrosion immune.

Main features of A350, taken from a June 2014 Airbus presentation

Main features of A350, taken from a June 2014 Airbus presentation.

The A350 and the 787 share many design solutions, such as CFRP based structure, wing design with high lift arrangement (cooperating spoilers and flaps for simple and effective high lift trailing edge, adaptive wing for optimal cruise lift versus drag ratio and load alleviation) and way of attaching the main landing gear by means of a Y spar arrangement on the inner trailing edge (the so-called Yehudi). They also have similar avionics concepts, based on networks of computing nodes connected with an adapted Ethernet network (AFDX).

The biggest difference between the A350 and the 787 is the way they deliverer power to the aircraft’s different systems. Boeing decided 2003 that it was time to go for an all electrical power conversion chain from the engine source to the power consuming system whereas Airbus decided to stay with a conventional system (inherited from the A380) which split the power delivery between an electrical conversion chain and a pneumatic one. While Airbus could migrate a modern classical system and its vendors to the requirements of the A350, Boeing had to start from scratch, architecting numerous firsts in the process such as water-cooled conversion electronics in an extra equipment bay behind the main landing gear bay. New designs from scratch means new challenges and Boeing had its share of operational problems with the more electrical systems architecture (conversion panel short-circuits, battery problems….). The all electrical architecture also means higher requirements on auxiliary power, the high electrical requirements of the 787 have made more than one airport power aggregate look bleak.

In contrast, the A350 system side has been smooth sailing with a high system maturity evident in the unusually trouble-free flight test period. This should also give initial airliner service an easier start. Once through the initial problems the more efficient power conversion of an electrical system should gain the 787 a slight efficiency advantage on the system side. In retrospect Boeing probably ask itself if it was worth all the hard lessons.

The A350 continues the Airbus tradition of a Fly-By-Wire aircraft with a family look-and-use cockpit (including side-stick), enabling an aircraft characteristics and cockpit commonality which has reduced aircrew cross-certification requirements since the A320. The Fly-By-Wire system is now in its fourth major revision with the signaling to the rudders going full digital; its maturity and adaptability has been obvious. We met with some of the A350 test pilots at Farnborough and they all said the A350 was very similar to an A330 to fly and the flight testing of the base characteristics of the aircraft had been virtually trouble-free.

System and Flight testing
Airbus has had the advantage to work with mature system concepts and their retained suppliers from previous programs. While all components have been brought up to date there has been no major architectural change in critical systems like flight control, power distribution, hydraulics, air conditioning etc. This has laid the foundation for a system and flight test that could focus on system maturity at EIS instead of getting the systems to work correctly. The final flight testing had been preceded by an extensive system test of individual components up to complete system rigs. But even before A350 hardware was available the correct functioning and interaction of all the myriad of system had been tested with digital models simulating the interaction of all critical systems.

As a consequence of all these preparations flight testing has been unusually trouble-free. The flight test team recognized they got an unusually mature aircraft when preparing for the 14 June 2013 first flight. This impression has since not changed. The aircraft has gone through flight testing without any major findings, allowing Airbus to stick to the 3Q2014 certification schedule set after previous development delays.

Program management
Airbus had major deficiencies in the program management of its previous program, the A380. The company was still suffering from having been a consortium of independent national companies, each convinced that they knew best how to design and produce their part, to the extent that they even reserved the right to have their own tools (the famous Catia version 4 to version 5 incompatibility between the German and the French parts being only the tip of the iceberg). Lessons learned from the A380 (and from observing the tortured development of the Boeing 787) were applied to the A350 program, not in the least in the company structure and program management. The company was galvanized into one by several restructuring programs (Power8, Future EADS) and program management was reinforced.

The A350 program manager, Didier Evrard, reported to the COO, Fabrice Brégier, and when Brégier was elected CEO, Evrard moved up as well in the corporate structure. Evrard thereby had the position to escalate any resource conflict or hold-up in the company. Further lessons learned were drawn from the troubled A380 program. All parts of the company were forced to work on the same toolset (the main being the product lifecycle management system, Windchill, and the 3D design tool, CATIA). Strict rules were introduced where every engineer and partner (in total 5,000 people including 60 partners) had to manifest every change of a design in the gigantic Windchill database. Catia was then used to visualize in 3D any changes in the design of the aircraft in real time. This visual mode, which is constantly updated to stay current, is called DMU (Digital Mock Up). The realtime DMU meant there was only one copy of the virtual A350 and one could rely on all the data in the database being current. This was a big change from the A380 where multiple copies of the aircraft were being worked on in the different countries and no-one knew which model was up to date and if so in what part of the model.

Another area which was given much attention was program management rules. In earlier projects, should, for instance, a part of the flap system not be finished for a critical design review, one could still let the wing pass the review and let the failing part catch up on a side-track to the main design. This rule, to keep the time plan and fix later, lead to the traveled work on the A380 and 787. It meant that work was being done by persons other than those who should normally handle the matter, requiring re-learning for these persons once the traveled work package was closed. It created inefficiencies and that knowledge was built up in the wrong hands. When finally the part was to be installed, one often had to disassemble the aircraft first as the part was now out of sequence. Further, the original tasks that should have been performed on the parts removed now had to wait until things are reassembled. Finally the knowledge had to be transferred to those in the production chain who should normally handle the task. Airbus changes this to the principle of “Stop and Fix” for the A350. It is better to take a delay immediately than later. This ensured that whatever problem caused the part to not meet the deadline the fix was worked on by the relevant persons or groups. It also builds the pre-production competence in the right places. This principle was also extended to production. Airbus was extremely cautious in the ramp of the test aircraft and later the first production aircraft, with the benefit of not having (too many) parts out of sequence.

The principle of “stop and fix” could be seen in work in the early parts of the program. The new composite construction caused some new problems to solve. CFRP parts don’t conduct electricity so an extra layer of metal mesh had to be introduced to serve as a lightning conduction part and as an electrical return path. The design of this rather complex “electrical structural network” took longer than planned. Another aspect of CFRP is that it is very strong in the fiber’s long axis but not in the traverse axis. Consequently a lot of work and testing had to be done on the so called “damage tolerance” of the aircraft’s skins. The strength of CFRP allows these skins to be very thin in places but the normal bumping into the aircraft by ground vehicles, docking for catering or cargo loading, could cause non-visible damage to the layered matrix. Hence the CFRP material thickness had to be increased beyond what was required from a structural point, as this adds weight the investigations into just how much were important and these took longer then planned.

The effects can be seen in the schedule graph which is a cut from Airbus presentations in early 2010 and late 2011. By 2010 these delays had eaten up all margins and the subsequent discovery of other issues lead to the rescheduling of the whole program, first to mid 2014 delivery, then finally to “second half of 2014”.

Airbus publicized program schedules from early 2010 and late 2011

Airbus publicized program schedules from early 2010 and late 2011.

These are just examples of the myriad of causes that forced Airbus to delay the program first by a year and then finally by 18 months. The effect of this principle was the there were delays in the initial and mid-part of the program. Once the necessary maturity had been reached for the different parts and systems, the time-plan stabilized and the test period was one of the smoothest ever. There has been no further delay since summer 2013.

Program communication
The A350 program has also broken new ground in how communication has been handled. Initially the communication tone was “we will not be delayed like the 787” i.e. a self-assured style like its competitor, with exact milestones that would not be touched. When it was seen that delays were inevitable, the communication tone changed and became more understated and cautious. Milestones were always given as “second half of 2013” rather than exact and it was always stressed that the program was very challenging.

Nowhere was the communication more elegantly handled then around the considerable weight creep the program experienced. All-in–all, the Operating Empty Weight increase from launch to EIS was in the region of 6%-7% compared with 12% for the weight-troubled 787-8. Boeing subsequently got that down to 9%. Where the outcry re the 787 weight creep was massive, it was virtually non-existent for the A350. Airbus waited with all hints of major weight changes until flight testing had shown exactly where they were, then the range was quietly cut with 350 nautical miles despite a change to a lighter two class interior of 315 seats (instead of the original 314 seat 3-class interior) and the addition of a maximum MTOW variant that raised MTOW from 268t to 275t. The few that noticed the range changes were told that if was indeed no specification change, the cabin had changed and so did the range. These changes represent more than an hour’s worth of fuel but few seems to have noticed (other than Leeham News, which reported it at the time). There has been no media attention on this blip. (The customers have, of course, known all the time; the only one who has been vocal about it is Emirates Airlines, which canceled their order for this reason among others).

This media strategy has rendered Airbus a lot of good will, to the degree that when slip up occur it does not have much news value.

Entry into service and production ramp
The challenge now for Airbus is the first customer’s experience of the aircraft’s reliability and how Airbus can manage the ramp up of production. Perhaps unfortunately for Airbus, the first operator will be Qatar Airways, whose CEO, the mercurial Akbar Al-Baker, isn’t the least bit shy about criticizing any of the OEMs on real and perceived shortcomings, often doing so publicly.

To ensure a high degree of visibility on the operationional maturity problems that can occur, the Airbus flight test operation worked as Airline number 1, i.e. all operational routines, etc. has been carried out as if the flight test was an airline operating the five test aircrafts. This has made it possible for Airbus to identify any problem areas that might not come up in traditional flight testing where aircrafts are fixed for glitches directly by the development organization.

Graph of "Airline 1" organization from Airbus June 2014 presentation

Graph of “Airline 1” organization from Airbus June 2014 presentation.

We have carefully monitored the production during recent years. The flight test aircraft had slips in their entry into service of typically 1-2 months over of what was originally presented. For the customer deliveries we see the chance of two aircraft delivered to Qatar this year and for a further six Qatar deliveries first Quarter 2015. This is enabled by the FAL start curve we see below presented by Airbus at the innovation days in June, the rate, presently at two starts per month, goes to three starts per month end of this year and these aircraft will be delivered end Q2 2015.

Final Assembly ramp of production taken from June 2014 Airbus presentation

Final Assembly ramp of production taken from June 2014 Airbus presentation.

Production time is around 8.5 calendar months for the first aircraft, then gradually reducing to seven months for deliveries in Q1 2015 and six months by next summer. Airbus has made several expansions in the production facilities in Toulouse and right now we see no reason that they should not deliver to their latest communicated plan.

A350-800 and -1000
The A350 is the latest in a range of clean sheet designs by the large OEMs (A380, 787, A350). Its fortunes will decide Airbus’ market share in the profitable large twin widebody segment. The smallest variant was officially canceled last week by CEO Brégier, its place being taken by the A330neo. This leaves the A350 production system free to focus on the above 300 seat segment where we are now waiting for its larger variant, the -1000. Programmed for mid-2017 EIS, its fortunes will be decided by how Airbus can control the weight increase versus the A350-900, right now by 10-13 tonnes. Its position was comfortable when the competition was the heavy and under-winged Boeing 777-300ER. This has all changed since Dubai Airshow last year where Boeing launched its answer to the A350-1000, the 777-8X and -9X. The -9X is especially troublesome for Airbus as it flies more passengers further at about the same seat cost as the A350-1000. For an airline that does not need the capacity, the A350-1000 is a safer choice with considerably lower trip costs. Should the airline need more capacity, the A350-1000 is pressed. We expect an answer from Airbus in the 400 seat segment in the coming years.

The A350 had a troubled birth, an initial development that generated delays and then a final two years where everything has gone to plan. The initial development tactic of accepting delays to build maturity in the design has paid off. The next phase of the A350 program will be no less interesting then the first, to follow the induction into service; it has potential to set new benchmarks for modern clean-sheet designs.

By Leeham Co EU

36 Comments on “A350 certified; we provide a comprehensive program review

  1. Excellent extensive program review !

    A good read.

    IMO the A350 is aimed at airlines looking for 300-350 seats long haul with a lot of cargo, with some growth baked in. If the 400 seat market is as large as Boeing hopes has to be seen IMO.

  2. Hello Leeham Co Eu
    Great day !
    Great piece

    We’ve already discussed the oddity of the 3 class to 2 class shift in the “typical Airbus” LOPA.
    It has been done on the A350 without any real reference concerning the “true” range of the final bird (including the overweight)
    The same as been done on the A330 family conducing to the same trend of decrease range .
    January 2012 detailed familly specs of all airbus :
    A330-200 238 t 3 class (12+36+205 = 253 ) = 7250 Nm
    A330-300 235 t 3 class (12+42+241 = 295 ) = 6000 Nm
    A350-900 268t 3 class (18+50+246) = 8100 Nm

    This as evolved in :

    A330-300 242 t 2 class (36+264 = 300) 6100 Nm (source LEAHY slides annual conf 2013, 01/2014).
    A350-900 268 t 2 class (48+267 = 315 PAX) 7750 Nm
    So on an existing airframe, with 7 t more MTOW, changing the reference LOPA induce a +100Nm range versus previous reference LOPA ?

    More digging : Airbus A330 family brochure 08/2014 :
    “The new take-off weight capability combined with the fuel capacity increase enables operators of these new A330-300s to carry additional payload on longer missions. Overall, the full payload range now increases by around 500nm over today’s 235 tonne A330-300, and by around 350nm over today’s 238 tonne A330-200.”
    6000 + 500 = 6100 ? something as gone wrong

    My take is that Airbus changed the weight of the furnishing to cope with real world 2 class LOPA and the widespread availabilty of IFE, Inflight connectivity, Crew rest and so on + real life catering. The OEW mentionned for A330 2 class LOPA is really close to real A330 LOPA’s.

    Airbus does the same for the A350… introducing a HEAVIER 2 Class standard LOPA decreasing the Range.

    They might have use this change to take the weight creep in account, but the new Lopa is heavier.

    It is using the same confort criteria as the standard 3 class A380 LOPA that is a true modern 3 class LOPA. We might discuss galley count maybe, but the fast is that some carriers are very close to the 525 typical PAX count event with premium heavy LOPA’s.

    Maybe we can find the truth by SAS who ordered A330-300 242 t @264PAX @5700 Nm and A350-900 @308PAX @7000 nm (SAS rules for range computation I guess)


    • Airbus giving “real numbers” for a very current setup ?
      Can we expect another long nose advertisement?

      • The A330 lost a bit design range over the last years because Airbus incorporated a heavier (more realistic) OEM cabin, increasing the OEW. That has nothing to do with advertisements.

  3. This is a fantastic achievement. Congratulations to Airbus! They were both cautious and vigilant at the same time. It paid off in the end. Champagne to all!

  4. It’s unfortunate about the range.

    Is it expected for later models of the A350-900 to become lighter and therefore regain an additional 250nm to 359nm in range? How does the current range compare to the 787-9?

    Does anyone know if the originally planned -900R using the wings, engines, and reinforcement of the -1000 will be launched?

    It is still listed on RR website and Wiki.

    • Significant weight changes will be introduced from batch 3 (MSN21 onward).

      Looking back at the 787, it took Boeing some 100 airplanes to get the weight under control.

      • More like 50. The first 50 were assembled with no eye for banalities like weight.

  5. Range and/or general Lopa’s/ seatcounts are not an issue for airlines.

    The amount of cargo they can carry on top of their full passenger cabin on typical flights from hot places in Asia is key.

    Big wings and engines rule these days.

  6. I have two questions for Leeham Co EU.

    Assuming the A350 is now certified to operate with Nickel-Cadmiun batteries:
    1) When will it be certified to operate with Lithium batteries?

    Assuming both types will be offered to all A350 operators:
    2) Which one between the two battery types is the most popular among the various customers?

    • Answer to 1:
      Airbus always maintained they would re-introduce Lithium-Ion batteries when the certification authorities had caught up and knew what requirements they would have re the certification of them. They are right now working with the authorities to establish the norms and will then reintroduce a conforming battery when these requirements have settled. Right now Airbus expect that to be 2016.

      2: The price of a (replacement) Li-Ion battery is higher but it lasts longer and it is lighter so it lowers the aircraft life-cycle cost (not in the least due to the lower weight) and that is the attraction to the customers. The A350 don’t really need the higher energy discharge capability of Li-Ion like the 787 does (to power the electrical brakes in a power out emergency stop), it should be pure economical reasons for ultimately going back to Li-Ion.

      • Thanks for you reply. Like your article it was very interesting and satisfying to read.

        Please note that the CSeries also operate with electric brakes like the 787, but is equipped with the traditional Ni-Cad batteries. The all-electric architecture of the 787 is more taxing for the batteries for various reasons, and Lithium batteries are therefore preferable. But my understanding is that they are not absolutely necessary. Battery power being much higher on the 787 the weight/space savings make Lithium much more convenient and economical. Plus the lower replacement cycle that you also mentioned.

        On the A350 the lithium batteries will be really small and very light. And in addition to their natural weight/space savings they will be made even smaller and lighter by sharing the same task between two separate units. In other words they will be doubled-up. So if the original plan has not changed I expect the A350 to have four smaller Lithium batteries instead of two larger ones like on the 787.

      • The 787 Li-Ion Batteries had extremely short service life even without catastrophic failure.
        Did this actually change/improve after optimisations during the grounding ?

        • The short life cycle was due to reliability problems. And these in turn were caused by design ans manufacturing problems. But all this has changed since Boeing decided to tackle the problem like if it was a life and death situation. Which it was in a sense.

          Therefore the Dreamliner batterie life should now be close to the kind of life cycle that was originally expected of this technology, but which had failed to materialize due to the above mentioned problems.

          The original concept was bad and still is. But at least it is now working normally. There is in fact little difference between the original design and the ameliorated one. But with that kind of power (fast and furious) it takes very little to create a thermal runaway situation as Boeing has learned the hard way during one of the darkest hour of its long and glorious history.

    • According to my knowledge the A350 is also certified with Lithium batteries because several of the test aircraft did use this type of batteries. The A380 was certified with the same type of Lithium batteries. As long as the EASA or FAA have no other rules Airbus has the right to certify the Lithium batteries according to the “old” standards.

      I guess Airbus feared FAA could change the rules right in the middle of the certification process and certified both systems.

  7. Airbus have stated that 3 flight test aircraft will be sold, although it will have no economic effect. (already planned for, I suppose) How long before they are delivered and have they been included in these estimates?

  8. During the last 15 years, Airbus engineering put 3 large innovative platforms into production; the A380, A400M and A350. The A330 and A320 NEO’s and A350-1000 will be relatively easy derivatives with existing supply chains.

    I, and many with me, have been wondering if this now mature aircraft development machine will scale down .. or role into a next project. A big twin, as some suggested, seems risky with questionable overall ROI, for Airbus.

    The other more attractive, larger gap in the industry is IMO the middle of market. Inbetween the A321 and A330/787 (that both are twice as big/expensive).Ranges up to, but no more then 5000NM. Super efficient, quiet in different sizes/ cargo variants. Replacing types like A300, A310, 767, 757, Tu154. The A330 and 787 are basically too big, heavy and expensive for those markets and only used because airlines have no alternatives. Marketsize? 2000-3000 in the next 20 years. Boeing will likely dismiss this segment because it overlaps partly with the new 787.

    • Hard to see Airbus scaling down their development machine, it cost so much to build. I guess COMAC would love them to, for sure anyone laid off would end with the competition.

      Next project is a big question, in my mind a big danger to the 777-X if Airbus go that direction.

      • It is from John Leahy. As a sales guy he’s not going to launch or talk about an aircraft with his customers before his engineers tell it it is possible. Until they make the decision on whether to go ahead or not, he has to convince his customers that the -1000 is adequate vs. the 777X. It was the same thing with the A330neo which he always maintained as not possible until close to launch.

    • Seems like a prime spot. Airbus just built a 65m wing, so they have that covered. The A320neo seems to be holding the 36m wing position for now. That leaves either trump the 777x with a 80m wing on a new design, or go with the 52m wing on a new fuselage width or use that wing on both a single aisle and double aisle?

    • great idea indeed. there were some rumors around that once the 320neos are out the next project there might be a new wing for the 321 to make it a true long range plane – current wing is actually undersized – and then based on that another stretch for short to medium range.

      I am in no position to calculate the economics of that plane or to understand how much you can technically do with a 30 year old design.

  9. Re: the “heavy and under-winged” 777-300ER, if it’s so undesirable then why has Boeing sold over 500 of them?

    • It is just a way ro say that this aircraft was excellent … and , even in the eyes of B, is somewhat outdated and justify a new design

      • Indeed, there was nothing derogatory meant to the 777-300ER with this line. Fact: it does have a small wing for its possible TOW, the wingloading is the highest in the business and the initial FL is low, the starting distance long at MTOW and acceptable only by virtue of really strong engines. There is no doubt the 777-300ER is a truly great aircraft, go back to where I wrote about how it exceeded all expectations in 2004. But now is 2014 and soon 2017 and then there is something better available, Boeing has also reacted in a convincing way (777X).

        • You could have said all this to begin with, instead of taking what sounded to me and most of my colleagues as a gratuitous slap at the -300ER.

        • I agree. I did not see leeham’s comment as derogatory either. Most feel the 300-ER is a great aircraft with a nicely balanced capacity, range, and fuel efficiency. In comparison to the same sized A350-1000, however, the 300ER will be heavy, less efficient, and relatively underwinged. The 9-x and 8-x will be even heavier but will have larger wings and improved engines. They will probably address different missions.

          To Leeham: do you know if the ultra- long range (9500 nm) A350-900R using the wings and engines if the -1000 will ever be launched? It is described both on Wikipedia and the Rolls Royce website when describing the Trent XWB for the -1000. Like the 8-x, It would make an interesting aircraft.

  10. apropos is there any specific reason around why the FAA significantly lags behind in certifying the A350 while the 787-8 appears to have got joint certification from both sides of the Atlantic. ( remembering something about crosswise complimentary certification arrangements ?)

Leave a Reply

Your email address will not be published. Required fields are marked *