Thursday, December 5, 2019

Media hand-wringing versus reality: aircraft test edition

Last week there were alarming headlines when the fuselage of a new model of Boeing's 777 airliner split while undergoing a stress test.

Boeing got an unexpected jolt in September when engineers in Everett put the new 777X airframe through an extreme test of its structural strength. Just as the test approached its target stress level, an explosive depressurization tore through the fuselage.

Boeing has kept the details secret, but photos obtained by the Seattle Times show that the extent of the damage was greater than previously disclosed and earlier reports were wrong about crucial details.

The test plane is a complete write-off, its fuselage skin ripped wide open just behind the wing. A passenger door that blew out and fell to the factory floor was a secondary impact of the initial rupture, which was located far below the door.

The relatively good news for Boeing is that because the test failed so explosively at just 1% shy of meeting federal requirements, it will almost certainly not have to do a retest. Regulators will likely allow it to prove by analysis that it’s enough to reinforce the fuselage in the localized area where it failed.

There's more at the link, including details of the test that led to the failure and photographs of the damage.  Interesting reading for aviation and engineering buffs.

The thing is, such testing is normal (as the linked article points out in smaller print, further down the text).  To suffer such a catastrophic failure is a nasty surprise to those nearby, but that's what the test is designed to induce, if necessary.  The fuselage and wings were already 99% of the way to successful completion of the test when the failure occurred, at a point far past the stresses that could conceivably be experienced during normal flight.  That's why Boeing appears unruffled by the failure, and will probably be allowed to demonstrate via calculation that it's added enough strengthening to affected areas to gain that extra 1% of safety margin.

The news media simply don't have the technical expertise any longer to adequately evaluate such events, largely because they don't want (or can't afford) to spend the money to employ specialist journalists who can understand and interpret them correctly.  I can remember when an "aviation correspondent" was actually qualified for his position in any one of a number of ways - former military pilot, former airline pilot, engineering background, whatever.  They knew whereof they spoke, and therefore they spoke with some authority.  Anyone from those backgrounds would have understood that this "failure" was, in fact, a deliberately planned experiment, and the "problem" was, in fact, a demonstration of considerable success.

Nowadays, in aviation as in almost every other field of activity, the journalist(s) concerned are likely to be media studies graduates, "full of sound and fury, signifying nothing", as the Bard would put it.  They don't understand the subject, so they mistake a legitimate engineering test (and its result) for a disaster.  It wasn't.

Which leads us to consider their political reporting journalism propaganda . . . but let's not go there.



Brad_in_IL said...

Good morning. There's an aviation-related YouTube channel, Mentour Pilot, who does all manner of videos about modern flight, planes, systems, etc. The channel host is a 737 instructor pilot from Sweden, and seems to know his shit very well. He just posted a description of the test, the 99%, the modeling & strengthening for the extra 1%, how the FAA approval should work, etc. His channel has a lot of good content, for those interested in the subject.

Uncle Lar said...

I recall one of the Spacelab missions I sat experiment ground support for was being covered by our local paper. Being a team lead I had intentionally scheduled myself to work night shift so would get home about the time the morning paper arrived so would give it a glance while unwinding before bed.
Late in the mission they printed a column headline, so-and-so experiment forced to shut down.
Terrible news, right?
One had to read far down the page to discover that said experiment shut down because it had processed all of it's samples and finally depleted it's battery power. And by the way gotten 150% of anticipated results as we had managed to run not only all the scheduled samples, but all the backups and spares brought along as just in case. Matter of a different experiment failing early on freeing up power and crew time, not an uncommon event on most of those missions.
This did destroy any remaining faith I might have had in the veracity of newsprint or any other mainstream media for that matter.

HMS Defiant said...

One of the many reasons the B-52 is still flying combat missions today, 60 years later, is that it was overbuilt by guys with slipsticks before computers and CAD. Me? I'd prefer a plane that was slightly overbuilt to meet the dynamic and static tests to one that was carefully computed to just barely scrape by the test on its way to it's 100,000 hour test with me or my daughter inside. Ify the design failed on a brand new air frame on the test bed it will sure as Hell fail after it's 50th landing or take-off when some other part of the plane gives way. We've had enough passenger deaths by sudden air frame failure.

Roy said...

It surprises me that some of you still believe *anything* put out by today's "journalists".

They lie.

They lie when they don't even have to lie.

They lie about everything - including the weather.

Don't believe me? What, exactly is a "heat index"? The way I see it, it is just another way for "journalists" to lie and make the conditions sound worse than they really are. Just look at how they react to a few snowflakes falling out of the sky in November.

The same is true here in this example. Tests fail all the time. That's why we have them, to find the weak points and take corrective action. Commercial aviation is safer today than it has ever been.

And comparing a BUFF to the average Boeing 777? Yeah. How many takeoffs and landings has the average B52 done in its 50 year career compared to your average, in service, 777 in just a couple of years?

Oh well.... "Gloom, despair, and agony on me!"

Beans said...

Well, trying to compare the B-52H (the version that is flying) to brand spanking new 777s is not fair.

Early B-52s suffered from too much wing flex, issues with cracking in the wing roots, all sorts of little issues. That got hammered out with each successive generation. Plus all B-52s flying today have gone through at least one major rebuild.

I seem to remember that the B-52 and 707 wing and fuselage prototypes were put through some serious stress and destructive testing.

The key in that sentence is 'destructive.' No matter how good the calculations, you need real-world testing.

Like, well, Airbus in the early years of its composite tails. How many cracks, breaks, losses of aircraft due to not stress-testing but relying on computer calculations.

Another plane, the A-10, is going through a re-winging because though testing and calculations showed the early generation wings were 'good enough', actual longterm use found issues with the wings and wing roots. Which are being addressed in the depot-level rebuilds during the A-10 life-extension program.

If we went by destructive testing only, we would all be scared to get in our cars, touch our computers, get within 100' of a water heater (300' if a gas water heater), etc., etc., etc. and, of course, etc.

Destructive testing has always been a valuable part of engineering and acquisition. How many things have not survived after being given to the acquisition team's pet gorillas? Toy companies have test teams of small household apes that are specifically tasked to destroy toys to see what happens. Military teams tend to give things, if they are smart, to a group of enlisted for the same reason, as enlisted have been found to be able to break anything (well, maybe not depleted uranium ammo, but maybe they tried...)

And always, the first rule of talking to the press is: DON'T TALK TO THE PRESS.

Second rule: The press will always misquote what is said.

Third rule: The press will always find the most stupid or freakish (or both) person to interview.

Fourth rule: If you didn't pay attention to the first rule, you're screwed.

Roy said...

Beans, I agree with everything you said except that first line.

Maybe I wasn't clear enough, but I wasn't comparing the B-52 to a "brand new" 777. The 777 has been in service for over 25 years. The point I was trying to make is that the comparison between the two vs longevity, is not valid. How many takeoffs and landings will the average BUFF make in a year compared to the yearly average for an in service 777.

I love your rules for talking to the press. I've been saying for years that an office building full of lawyers and accountants could burn down, and the man-on-the-street the press will find to interview will have one tooth and not be able to formulate a complete sentence.

Silent Draco said...

Roy and Beans made my tester's heart chuckle.

Operational T&E is a phrase to strike fear in every heart, and weakens a reliability engineer's bowels. Every calculation goes out the window, to include safety and durability. Story I got from one of the test team: The HMMWV failed spectacularly in original OT. Someone forgot to tell the troops that all-terrain didn't mean going over rocks the height of the suspension and differentials. After checking headspace and timing, retest showed reliability about two orders of magnitude higher for mean miles between OMF.

If it doesn't break, didn't test sufficiently and disaster waits. If it breaks too much, bad design or overtest. In between, probably on track.

HMS Defiant said...

I still recall that about the time a Minuteman blew up in the silo, a B52 was taxiing for take off after refueling and the wing just fell off. Stress is present in aircraft and to design one that fails to take that into account or to simply push it to side when the plane fails is kind of where we're at right now. Look no further than the 737 which are, still grounded because they chintzed the software AND the testing.
I wasn't comparing the B52 to the 777 but I'm willing to bet you won't find any of the latter still in the air 40 years from now.

dGarry39 said...

Another bullet that hit hard the Boeing engineers. They definitely get to up more the standards but keep the cost within the profiting level.

Roy said...

"I'm willing to bet you won't find any of the latter still in the air 40 years from now."

Really? Ya think? Well, the B-747 just recently retired from passenger service, and they are still going strong in freight service. How long ago were they placed into service? (Hint - 1970.) The B-737, that you have claimed is grounded is one of the most common and log-lived jets in commercial service - even longer than the 747 as it came into service in 1967. (*only* the variant known as the 737-800 Max had the software problems.)

But again, it's not really fair to compare a military plane like the B-52 with any commercial jet for various reasons. I could be wrong, but I would be willing to make a small wager that any average 10 year old B-737 has had more takeoffs and landings in that 10 years than a 50 year old B-52 has had in it's entire lifespan.

Larry said...

Early Buff vertical tails suffered a number f usually catastrophic failures, too.

Nuke Road Warrior said...

As a former test engineer, I am always amused by press articles decrying the failure of this or that system during testing. It's testing after all. You test because no matter how good the design engineers are, no matter how good the latest whiz-bang software is, you can't account for all the real world variables until you put the system to the test. Is every flaw going to be identified during testing, of course not. You hope your test program will uncover the most egregious and dangerous flaws and allow corrective action to be taken. No system is perfect out of the box, it was designed by humans, after all. Sometimes things sneak by the testing phase and people die. However, the more failures you uncover during testing, the fewer catastrophic failures you have in service.