By Leeham Co EU
Lessons learned by Airbus on its A380 production and development by arch-rival Boeing of the troubled 787 appear to be paying off with the A350 XWB.
There are now two A350s operational in the flight test program as it counts down to a fourth quarter delivery target for launch customer Qatar Airways. Testing has passed the 1,000 hour mark and by all accounts is going well. Three test aircraft are coming on line in the next four months to complete the 2,500 flight hours needed for certification. After 1,5 years of delays, the flight test program appears proceeding smoothly and tracking to plan.
Boeing and Bombardier should have had it so good with the 787 and CSeries. The 787 program was delayed nearly four years, interrupted by design and production issues and an in-flight fire on final approach to a landing in Texas involving a power control unit. Bombardier last month announced a new delay, its third, in the CSeries countdown to EIS, this time of 9-15 months.
With flight testing heading for certification in August-September, Airbus says the big challenge is now the production of the serial airplanes. Having been following the production preparations over the last two years, here is our view on how Airbus stands in their industrial ramp up. Airbus plans to ramp up to 10 A350s per month four years after EIS, and it is talking with suppliers about a higher rate.
The Airbus production system
Airbus has since its inception in 1966 used the principle of pre-equipped sections of the aircraft which are completed are remote locations and then flown to a Final Assembley Line (FAL) at Toulouse for final joining together to a complete aircraft. This principle, called FAL and pre-FAL, was invented by Airbus production father, Felix Kracht, as gifted as its original chief designer, Roger Beteille. The pre-FALs were a clever move to give the participating countries enough systems and integration work for them to give up their demand for multiple final assembly Lines, a devastating inefficiency that had plagued previous cooperation projects in Europe.
The historical pre-FALs have today morphed into competence centers with accompanying production facilities. The pre-FALs for the A350 are:
· St Nazaire on the French Atlantic coast, competence center and pre-FAL for forward sections and middle sections.
· Hamburg, competence center and pre-FAL for rear sections and vertical tails.
· Filton and Broughton UK, competence center and production unit for wings. These then gets shipped to Bremen who does the final pre-FAL assembly of the wings with their high lift systems.
· Getafe Spain, which is the pre-FAL for the horizontal tailplane produced at Puerto Real
· To this shall be added the engine, the Rolls Royce TXWB, which is produced at Derby UK.
All these pieces of the total airplane are produced as complete modules with all systems installed except cabin walls, seats, galleys etc. They are also tested for their correct functionality at the pre-FALs.
The Final Assembly Line, FAL
At the A350 final assembly line all these sections are built together at seven different stations. The stations are named in reverse order; the pre-FAL sections are gathered at Station 59 and final checkout of the finished aircraft is done at Station 20.
We have followed the A350s production through these stations over the last two years. The time that each aircraft spends at the FAL stations gives a good picture of the production readiness of the program because any non-maturity of the sections or traveled work means more work packages have to be completed at FAL and more time spent in the troublesome stations. The following graphs shows how the FAL has performed since the first test aircraft MNS001 entered FAL in July 2012.
The first graph shows elapsed time in the different stations. One can see how gradually the FAL time has gone down from 300 days to (our prognosis of) 220 days for the last prototype, MSN005. This prototype is now in Station30 and will be flying end of May:
Here are the stations and the operations performed in the stations:
· S59: the reception station for pre-FAL sections and where fundaments are fitted which are too large to enter through the doors (galleys, toilets, large fundaments)
· S50: the fuselage join station
· S40: the wing, horizontal and vertical tail join station where also pylons are hung and the cabin walls are fitted. During the operations the electrical system is also checked out for the completed aircraft.
· S30: Checking of all systems and installation of cabin seats.
· S18: External checking of pressurization, radio reception and tank checks.
· S20: Engines are mounted and final cockpit equipment is fitted.
· Test before EIS or delivery to flight test.
Airbus produced a video showing how the FAL functions
To make the FAL time graph more readable, we have added 20 respective 10 days of cabin installation days at S30 and S20 for MS004, which does not have a cabin nor a heavy instrumentation fit. Without this addition, it would have been difficult to see the trend for other MSN004 stations.
Table 2 is a production dates representation of Table 1. It has actual and postulated production dates in the different stations:
From these 2 charts one can draw some conclusions:
· The A350 FAL shortens the assembly times in the different stations monotonously. This shows an early learning curve and the diminishing level of elaborate test instrumentation installations. It also shows a low level of FAL dramas, i.e. traveled work or parts that does not fit.
· MSN5 is the first serial like aircraft with virtually no test instrumentation. Its FAL time is a good yardstick for the time MSN006, the first A350 for Qatar, will spend in FAL.
· With added test time for final inspection and delivery to Qatar, the predicted time for delivery is still ahead of the communicated times of December (Airbus) and November or October (Qatar).
From the above it is clear Airbus has built in margins for any delays in certification or delivery issues with the first EIS A350 for Qatar. One can also see that Airbus should be able to deliver an additional three aircraft before the year draws to a close if things run to plan.
Interesting reading Scott. Any idea how do the times compare with other Airbus aircraft? And any idea which of the sections – if any – is the focus of most time reduction now?
On another note, final passenger flight of the DC-10 today (http://www.bbc.co.uk/news/uk-england-birmingham-26259236). Bit of a whimper there really.
I’ve read MSN8 (SQ #1) will be loaded into the FAL after MSN11. Airbus build MSN’s out of sequence. Perhaps you can check this information with Airbus.
We think that could very well be. It would mean the first 5 series A350 (MSN6,7,9,10,11) would be all for Qatar before a new airline variant would enter the FAL, MSN008 for SQ. This would be sensible and would fit with the declarations from SQ that they will get their first A350 in 2015.
A350s will go for a second cold / hot weather tests in MAy, Indoor. I guess 1 such a WB facility is good enough for the world.
Wonder what the US Air Force charges to use that cold soak place in Florida?
Interesting report. What to make of the ‘Leeham Co EU’ byline? Have you added another analyst?
I mentioned some time ago that I now have a European affiliate.
Interesting article. QA could be improved a bit however:
“There are now has two A350s operational” (right at the beginning)
Would be interesting to see numbers of comparable programs like the 330 and 787.
I’m wondering where the 350 ranks in total build time.
Thanks. A little dyslexia at work.
And, after reading, I’m still wondering, what are the, “Lessons learned from A380, 787 [that] benefit [the] A350?”
Just a few: use the same CATIA version throughout the organization. Don’t give customers interior seating design carte blanche. Be much more careful designing hybrid carbon-fibre/metal structural components in weigh savings programs. Watch suppliers much more closely. Don’t specify fuselage fasteners that aren’t available. Be more careful using less experienced sub-suppliers. Get a handle on system nuisance warnings earlier in the test program. Keep a tight lid and carefully control information to the media from inception through product life.
IMU the Catia bruhaha was a bit more subtle than just ignoring SW version mismatch. The adaption layer introduced was faulty.
Don’t have a faction fight in management while doing your biggest project ever ( A380 )
Only hand out unfibbed information. Bad goo will seep out of the seams anyway.
Don’t have your PR department pose as engineering. ( B in general )
Believe your suppliers if they say that there is a nn month lead for items.
The leap improvement in processes & tools is not restricted to just using a same CATIA version, but a bunch of other common tools (accross what 14 years ago were different companies) from electrical design, to systems, configuration, planning, EVM…
Great post, Scott. I’m intrigued about the sources your European affiliate has.
Not specific to this piece but did you notice RR youtube:-
Thanks Scott. Excellent report and appreciate the detail. Glad someone in the A&D has been able to leverage lessons learned!!! Maybe, all will watch this approach and get some other programs on successful tracks!!!! About time
This is obviously only possible because Airbus isn’t pushing the A350 as much as Boeing did and are only relying on the engines, duh.
Is this a dig at me? Sorry for the delay Bryan but that may be true and it may be they did not push as hard in design approach. Look, two companies two approaches that end at the same point. If engines generate performance value then you use them. If on Boeing’s part, they failed to develop airframes over 20 years, they might have issues brining a product to market. Might explain why they dailed the 777X back. but you would know more than me.
That, “Airbus produced a video showing how the FAL functions,” spends more time looking at the roof than inside.
MSN5 was towed to station 18 this evening.
Someone spotted MSN5 at station 18: http://tinyurl.com/n74dvak
Some very good in depth detail, well done. Now if Av Week would return to its roots instead of the fluff they serve up these days………………
Still keep in mind that Boeing did not find the side join issue until late in structural testing.
Also missing through published in other sources previously is that the first 17 aircraft off the line are then modified 80% for the next series in structural (wing and body I believe). Most airlines don’t like to take that sort of one off copy.
That’s a massive change. So the supply entities are not going to get going as they have to regroup. It also unsettling to have that much engineering not done right. As with the A380 wing cracks, some of that could take time to reveal themselves.
I believe there was two or three steps (trenches as they say over there) after 17 aircraft before they hit the supposedly final production article.
Certainly no slack for any glitches and you have to wonder if they can ramp up.
On the other hand, Boeing has managed to single handed set such an awful standard that it is not going to be hard to beat it.
“Also missing through published in other sources previously is that the first 17 aircraft off the line are then modified 80% for the next series in structural (wing and body I believe). Most airlines don’t like to take that sort of one off copy.”
Apparently. Its in the news today.
“Also missing through published in other sources previously is that the first 17 aircraft off the line are then modified 80% for the next series in structural (wing and body I believe). Most airlines don’t like to take that sort of one off copy. ”
The change is 40%, not 80%. And it is mostly cabin related.
Can someone explain why S18 is longer for MSN9 than for the earlier planes?
Hmm, it isn’t longer in the “start normalised” graphic?
I think it would be the 2 week shutdown of the FAL during the industry holiday in France in August, you find the same thing on other stations for MSN6 to 9.
Pingback: Learning Curves (case in point) | The Blog by Javier
Pingback: Learning Curves: Boeing 787 case in 2013 | The Blog by Javier