Debugging the A350-XWB
February 14, 2014 8:47 AM   Subscribe

How Airbus is Debugging the A350-XWB. Jeff Wise, writing in Bloomberg Business Week, describes the 18-month testing process for the new Airbus A350-XWB passenger jet. One page version (printer format). And a bonus media offering: a somewhat functional online 3D flight simulator.

As a flight engineer and head of the department, du Ché gets first pick of the test flights. Although he describes himself as risk-averse, he tends to choose those he calls the most “interesting,” which means at the edge of the plane’s capabilities,

Bonus video of the A350-XWB's maiden flight.

Story.
posted by spitbull (30 comments total) 13 users marked this as a favorite
 
By pulling a lever, the crew can trigger a set of explosive charges that will blow a hole in the right side of the fuselage. They can then leap down a slide, through the hole, and into the air.

Huh. We have this on the plane I work on. They disabled them all years ago because they decided jumping out of the plane and breaking your back on the leading edge of the wing wasn't a great idea.
posted by backseatpilot at 8:54 AM on February 14, 2014 [2 favorites]


Is this really a flight simulator? I can't find any controls. +1 for a nice 3D view of the plane, though.
posted by JoeZydeco at 9:02 AM on February 14, 2014


As someone in software, 18 months of testing makes me so jealous.
posted by idiopath at 9:02 AM on February 14, 2014 [15 favorites]


We have this on the plane I work on.

Did yours have the slide? What's the best way to bail from a commercial airliner like this?

From an engineering standpoint, this plane is amazing. There are so many huge parts constructed from composite materials, not just a bunch of pieces of aluminum bolted/welded together.

Does anyone else find the first infographic, with the two planes flying a graph, to be incredibly frustrating? There's a yellow one and a black one, but there's not a legend to tell me what that means, and the graph doesn't stay up long enough for me to study it.
posted by spikeleemajortomdickandharryconnickjrmints at 9:03 AM on February 14, 2014 [1 favorite]


What's the best way to bail from a commercial airliner like this?

On the ground, with an inflatable slide.

You really can't just jump out at the earliest opportunity, unfortunately. When our system was operational, it could only be used below 10,000 feet and you had to depressurize the airplane before the hatch would open. You've also got to carry parachutes for everyone, which for a flight test probably isn't too big of a deal but for an operational aircraft is going to pose problems. We normally have a crew+passenger complement of around 30-40, and not only is that hundreds of pounds of parachute that isn't fuel you could be burning, having all those people scrambling trying to don parachutes in an emergency is going to get chaotic.

Once you're below 10,000 feet, you might as well try to land the thing. If it's really spiraling out of control, it's probably broken up at that point and if not then you've only got a handful of seconds at most before you get really intimate with the ground.
posted by backseatpilot at 9:10 AM on February 14, 2014 [8 favorites]


Hey, give it up for backseat pilot. Thanks for hopping in aviation threads. It's always interesting.
posted by spikeleemajortomdickandharryconnickjrmints at 9:12 AM on February 14, 2014 [3 favorites]


By pulling a lever, the crew can trigger a set of explosive charges that will blow a hole in the right side of the fuselage. They can then leap down a slide, through the hole, and into the air.

Huh. We have this on the plane I work on. They disabled them all years ago because they decided jumping out of the plane and breaking your back on the leading edge of the wing wasn't a great idea.


buhh...Explosive charges built into the fuselage of the plane? How did this idea ever get beyond someone's brain bucket? How typical a solution is this in aircraft aviation design?
posted by bird internet at 9:18 AM on February 14, 2014


How typical a solution is this in aircraft aviation design?

Well, it worked for the Mercury capsules. And if you don't like blowing a hole in the side of the plane you could always sit on a rocket motor.
posted by backseatpilot at 9:25 AM on February 14, 2014


How typical a solution is this in aircraft aviation design?

Explosive Bolts
posted by spikeleemajortomdickandharryconnickjrmints at 9:25 AM on February 14, 2014


Is this really a flight simulator? I can't find any controls.

Overheard during the Design Team meeting: "Hey, if they can't figure out how to disable the autopilot, we don't want them flying our plane."
posted by achrise at 9:28 AM on February 14, 2014


You also keep several explosives in your car, assuming it was built within the past decade or two.
posted by backseatpilot at 9:28 AM on February 14, 2014 [1 favorite]


> Explosive charges built into the fuselage of the plane?

Right on... The combination of "airplane" and "explosive" kind of unavoidably invokes mental images of large-scale charges like sticks of dynamite, but you should think of it more like a single-use device that irreversibly removes one thing from another thing with sudden force. That includes everything down to a small shell with just enough power in it to pop a bolt out of the pieces of metal it's joining.
posted by ardgedee at 9:36 AM on February 14, 2014 [1 favorite]


By pulling a lever, the crew can trigger a set of explosive charges that will blow a hole in the right side of the fuselage. They can then leap down a slide, through the hole, and into the air.

Scriptwriters, start your engines.
posted by longdaysjourney at 9:41 AM on February 14, 2014


Explosive charges built into the fuselage of the plane? How did this idea ever get beyond someone's brain bucket?

Ejector seats have been around for a while in fighter planes. The awesome thing is that early ones used a catapult.
posted by srboisvert at 9:45 AM on February 14, 2014


Explosive charges built into the fuselage of the plane?

Not just the fuselages. Look at fighter or attack aircraft sometime, and you'll see several types that have jaggedy-line-things in the "glass" of the canopy. They're not there for stealth -- that stuff, right over your head if you're flying it, is an explosive cord to shatter the canopy so you can eject through its debris.
posted by ROU_Xenophobe at 10:08 AM on February 14, 2014


As a test engineer, I'd honestly be looking at automating a lot of this testing for the next generation. Why put humans at risk at all? They're not going to be doing the piloting in the long run anyway.
posted by underflow at 10:14 AM on February 14, 2014


"Hey, if they can't figure out how to disable the autopilot, we don't want them flying our plane."

I thought Airbus pilots pretty much fly the plane via the autopilot. I was just watching this video of an A380 landing at SFO and the approach is 10 minutes of the pilot turning the little heading and altitude knobs every time ATC requested a change.
posted by JoeZydeco at 10:14 AM on February 14, 2014 [2 favorites]


Scriptwriters, start your engines.

On the plane ... Mandy leaves her seat and stumbles to the rear of the plane, where she suddenly kills one of the flight attendants with a syringe.... Mandy lines the back exit with explosives before putting on a protective jumpsuit and parachute. She retrieves an explosive device hidden in a fire extinguisher and sets the timer on it for thirty seconds before blowing open the back door and ejecting herself from the plane. Seconds later, the explosive goes off, destroying the plane.

-- Day 1: 12:00am-1:00am
posted by dhartung at 10:35 AM on February 14, 2014


this video of an A380 landing at SFO

I love seeing professional flight crews in action, all the training leads to such a calm cockpit, and clear communication. I wasn't expecting to watch the whole video, but it was so relaxing.
posted by Harpocrates at 10:39 AM on February 14, 2014 [1 favorite]


As a test engineer, I'd honestly be looking at automating a lot of this testing for the next generation. Why put humans at risk at all? They're not going to be doing the piloting in the long run anyway

One thing about humans is that they're pretty good at noticing the unexpected and the small details, at least some of the time. I think in this day and age, by the time the first plane is sitting on the tarmac, the odds of it being a new Comet are pretty small.

(It's really too bad about the Comet, it was a lovely looking airplane and quite advanced for its day. Minor details like breaking up in mid-air aside...)
posted by maxwelton at 10:43 AM on February 14, 2014


As someone in software, 18 months of testing makes me so jealous.

As someone who builds websites on stupid-short timelines, any "pre-launch" testing at all makes me jealous.
posted by maxwelton at 10:44 AM on February 14, 2014 [2 favorites]


How Airbus is Debugging the A350-XWB

Thoroughly, I'd hope.
posted by Aizkolari at 11:00 AM on February 14, 2014


I can control the animated plane with my trackpad -- Chrome, OSX, HTML5 and Flash enabled. It's pretty crude.
posted by spitbull at 12:05 PM on February 14, 2014


If it ain't Boeing, I ain't going.
posted by ocschwar at 12:30 PM on February 14, 2014 [4 favorites]


(It's really too bad about the Comet, it was a lovely looking airplane and quite advanced for its day. Minor details like breaking up in mid-air aside...)

Don't feel too bad for it - it had a good alternate existence as the Nimrod.
posted by ROU_Xenophobe at 12:32 PM on February 14, 2014


Nimrod had its own issues.
posted by backseatpilot at 12:59 PM on February 14, 2014


oschwar- even tho i'm a born and raised Puget Sounder: don't be too certain about that. lots of the stuff they're shipping up from South Carolina is chock full of crap execution/bonehead issues and management is doing all it can to get work away from the unions in everett. not to mention the problems they've had with the 787 since rollout.
posted by cult_url_bias at 1:57 PM on February 14, 2014


cult_url_bias:

Too true. The recent management moves are deplorable. But Boing's slow and cautions approach to adding software avionic controls is the right one. Airline pilots should never have to call tech support to land their planes.
posted by ocschwar at 2:27 PM on February 14, 2014


As someone in software, 18 months of testing makes me so jealous.

If they make a mistake, a plane full of people could die. Makes me not so jealous.
posted by Soupisgoodfood at 11:34 PM on February 15, 2014


> The A350 test flight program so far has uncovered half as many problems as the A380’s did—evidence, du Ché says, that Airbus’s debugging strategy is working.

I assume the reporter shortened a longer explanation here because as-written that metric is not sufficient to demonstrate the plane is safer than the A380. The crucial missing element is demonstrating that the known-vs-unknown defect ratio of the A350 is proportional to that of the A380.

As an engineer, let me tell you: the statement "we found fewer defects" is, first, a cause for deep concern, and only later a cause for much rejoicing.
posted by sixohsix at 2:26 PM on February 16, 2014


« Older No spring can follow past meridian   |   Don't Talk Newer »


This thread has been archived and is closed to new comments