Materials Failure in Airplanes - Chardok, Keep Out!

[Broomstick]

:::steps to podium at front of room, looks out nervously over audience:::

"Hi, my name is Broomstick and I'm a disaster groupie."

:::group chorus::: "Welcome Broomstick!"

But seriously folks, I've had a fascination with disasters of all sorts for a long time now. I have a particular interest in aviation disasters, of course, due to personal interests and bias. Anyhow, media reports frequently speak of "metal fatigue" and "stress fractures", but how often do you see them on the nightly news? About as often as morgue photos, I'd say. Well, here's a couple pictures for your education and edification:


This is what the interior of an airplane wing looks like after it snaps off in flight (and spends a couple days in the Carribean). That occurance is what is known as a Bad Thing. Here is the handy caption provided by the NTSB in their usual dry manner:

"Overall view of the inboard end of the right wing as it was recovered from the water. An unlabeled arrow indicates the lower spar cap of the rear spar where fatigue features were observed."

Note that most of the interior of the wing is empty space. In this particular model of airplane, in actual operation much of that empty space is filled with fuel. This is a "wet wing" where the wing interior is sealed and serves as the fuel resevoir rather than having a separate and discreet tank. Note also the jagged metal where the non-fatigued structures and skin tore after the wing spar failed. Note the wiring dangling, some of it electrical in nature. No doubt when the fuel-containing wing snapped there were sparks. This is why the airplane caught fire after the fracture.

Oh, yes - it's not obvious, but in the above picture the wing section is actually upside down. That is, the upper surface in this photo is, in normal operations, the bottom of the wing.


A close up of the actual fracture point of the spar cap


"Closer view of the fatigue region in the horizontal leg of the lower spar cap of the rear spar. Unlabeled brackets indicate fatigue origin areas at the surfaces of the fastener hole, and dashed lines indicate the extent of the fatigue region visible on the fracture surface. "

A further note - this failure occured in a 60 year old airplane operated off salt water. It is not known at this time how much corrosion may or may not have contributed to this accident. It is not known how much age may have contributed to this accident - it is possible this spar cap is younger than 60 years (even major parts may be replaced in an airplane that old).

Just to be fair, it isn't always metal airplanes that have problems. Here are some views of compression fractures in wooden wing spars (pre-catastrophic failure - these were caught by inspection and the spars removed before a Bad Thing occured, which is how the system is supposed to work):



rear of same spar:



Something very subtle that may have zipped past you - this spar contains a "grain anomaly". Wood used in airplane contrastrution, needless to say, needs to be as flawless as possible. That calls for absolutely straight grain. The defect is most obvious in the rear view, as a wave in the grain to the right of the fracture.

Here's a top view of the same spar and fracture:





Another spar crack:



Here's a bad one - the spar has actually separated. The last time this airplane landed the only thing holding the wing together was the cloth skin. Which gives me new respect for dacron, but YIIIIIIIE! This spar is now just very expensive firewood.

[Chardok]

*pokes head in*

.....



...........


......................


TOLDYASO!

*runs away scraming madly*

[Mrs Kendall]

Aww, Chardok! You weren't supossed to come in here You bad boy!


*hugs Chardok and makes him feel happy again

[Broomstick]

:::makes note::: Chardok seems to be sort of person who, upon being told "don't do that", feels overwelming compulsion to do exactly that...

Those cracks in the wooden spars, by the way, are why the regs require parachutes while flying aerobatics. Also why, if your g-meter records an over-stress situation you MUST proceed slowly and gently back to the airport and have the spar inspected prior to flying again, even if everything seems normal.

Wood fails differently than metal, in that cracks from either fatigue or stress are more easily seen in wood than metal. The DuPage Flight Standards District Office used to have a cracked metal prop in the lobby as an educational display. Even if you knew what you were looking for, and where it was located, it was still difficult to see the crack in the metal. Compare that to the wood spars above, where the cracks are much more apparent.

[Darth Wong]

Metal fatigue cracks in compression are much worse, because you can't even see them upon close inspection. In one case an oil rig went down because one of its support members failed due to fatigue despite having been recently inspected for fatigue cracks with ultrasound equipment.

[RogueIce]

Metal fatigue cracks in compression are much worse, because you can't even see them upon close inspection. In one case an oil rig went down because one of its support members failed due to fatigue despite having been recently inspected for fatigue cracks with ultrasound equipment.

So is there any way to spot them? Some kind of special method or something?

[Broomstick]

Not sure if I'm going to express this well, but here goes...

One of my chronic irritations (likely one of Mr. Wong's as well) are wankers who talk about how miracle material XYZ is going to render metal or plastic or something else obsolete. I have some understanding of why this is not so, but since I'm not an engineer it's hard for me to explain to others, especially the above referenced wankers.

This, however, is an example of why one material does not necessarially replace another. There is no question that metal is stronger than wood, and since metal is manufactured rather than grown the quality is much more consistent. These are just two reasons that metal is used to build airplanes.

We do, however, continue to build wooden airplanes. In the OP pictures, the metal airplane is 60 years old. The wooden spars shown are from airplanes "only" 20-30 years old. Why are we building wooden airplanes still? Also, I did not mention it, but these spars are from Bellanca airplanes - things like Decathalons and the Citabria I've been flying this year. Citabrias are built to withstand 5.5 g's on a routine and regular basis, Decathalons 6 g's (there's a 50% safety factor as well, so the wings shouldn't break off until 8-9 g's) Wow, that's a lot, isn't it? Why don't we build those planes out of much stronger metal?

In part, it's the failure mode.

As Mike pointed out, locating a compression fracture in metal isn't going to happen with naked eye. It can be difficult even with tools designed for the job. In fact, the metal airplane in the first post passed its most recent inspection for spar damage.

In the case of the wooden spar, the fracture is visible and identifiable by the naked eye even before catastrophic failure. If I, as a pilot, pull off an inspection plate on the wing and look inside and see that kind of spar damage I know not to fly - with no tool more sophisticated than a screwdriver required. If it's a metal spar there's no way I'll see that accident-about-to-happen. Again, note that the problem with the wooden spars was detected BEFORE the accident - the crack in the metal airplane eventually killed 20 people.

So... do you choose a material of less strength and durability, but with a slow failure and easily detected failure mode ? Or one less likely to fail, but also less likely to warn you of failure?

The equipment required to detect incipient metal failures of this sort is expensive. Aerobatic aircraft are regularly subjected to high g forces (even "normal" category aircraft are required to withstand 2g's for prolonged periods of time). For an amateur-level aerobat, weekend flyer, private owner, and so forth having the slow-failure mode makes it much more likely that the cracks will be detected - since inspection is easier and cheaper it occurs much more frequently. And that's why a LOT of aerobatic aircraft are still made out of wood. It's an easy matter to whip off the inspection plates after a flight or competition and look for potential trouble. Bush planes flying to remote areas also are frequently made out of wood - again, it's an issue of inspection. If you're in a remote area you just don't have access to sophisticated testing equipment.

And it really does come down the inspections - material costs for metal vs. wood planes on the scale I fly come out to more or less the same.

This issue also arises with composites. Composites can be as strong and consistent as metal, but with the lightness and flexibility of wood.. They're fantastic for aircraft construction.... except for the failure modes. At present, it is still more difficult to detect flaws in composites than in metals. They do not give warning, at least not in a way detectable to bare human senses. Certainly, there are people working on this problem. But, again, despite some folks saying for years that composites would replaced metal in aircraft they have not. They are certainly replacing some metal, but not all. Nor do I think they ever entirely will. There may come a day when composites are the most common aircraft building material, but just as we continue to make wooden airplanes for good reasons, we will continue to have a niche for metal ones as well.

[Broomstick]

Metal fatigue cracks in compression are much worse, because you can't even see them upon close inspection. In one case an oil rig went down because one of its support members failed due to fatigue despite having been recently inspected for fatigue cracks with ultrasound equipment.

So is there any way to spot them? Some kind of special method or something?

Wong might have more to say about that, since he is an engineer. I know this strictly from a pilot's perspective.

Yes, there are devices to detect fatigue and cracks in metal. The testing is expensive. It often requires parts to be shipped to another location for testing. One of the reasons helicopters cost three times as much to maintain as fixed-wings of comparable performance/size/capacity is becuase periodically the rotor assemblies must be removed, disassembled, and shipped out for testing. Failure to keep up with this can result in the helicopter part suffering catastrophic failure in flight. This almost always results in helicopter bits flying off in all directions at high speed and the death of all occupants as their various pieces (yes, it can be that violent) plummet towards the ground.

Even so, the testing for this sort of damage in metal isn't perfect and problems are sometimes missed.

[Ace Pace]

*reads* So basicly they are never really sure if a metal airplane is in good shape?

Very informative read.

[Alan Bolte]

Hmmm. I wonder how one might create a composite that more obviously indicates materials failure without compromising its strength. Perhaps it might be possible to create some sort of surface coating which is sensitive to microscopic fracture and creates a visible cue. Of course, this is just wild speculation of the sort that can at best lead to research.

[Broomstick]

*reads* So basicly they are never really sure if a metal airplane is in good shape?

Strictly speaking, you're never 100% sure ANY aircraft is in good shape - that's why we practice procedures for various emergencies.

Hmmm. I wonder how one might create a composite that more obviously indicates materials failure without compromising its strength. Perhaps it might be possible to create some sort of surface coating which is sensitive to microscopic fracture and creates a visible cue. Of course, this is just wild speculation of the sort that can at best lead to research.

That's not a bad idea, however....

With composites one of the possible scenarios is not a surface crack propagating into something larger, but interior flaws or even voids that show no visible sign on the surface.

I'm going to try to find an example of composite failure, be back soon...

[Broomstick]

This is an example of aircraft composite failure. Specifically, this is what is left of the composite rudder on an Airbus passenger jet.

I will also mention that no human being was hurt during this incident - although, as you can see, all but a fragment of rudder was lost, the pilots were able to land safely. Remember what I said about practicing for various emergencies? Having a backup plan is important.

Regrettably, that was not an isolated incident. This is a picture of the composite tail from the Airbus involved in the Flight 587 accident in New York City in November of 2001 which resulted in the death of all aboard:



I think part of the problem is that while we are familar with the failure modes of wood and metal, composites are still new enough that we aren't as knowledgable about either their failures or how to detect them.

Now, for extra bonus points - anyone care to hazard a guess as to why two very similar accidents (complete loss of substantial part of tail in flight) had such different outcomes? Remember, in one case everyone died. In the other, nobody was even hurt.

[LadyTevar]

It looks like in the top photo, only the moving parts of the tail fell off, while in the second the entire tail did.

Am I right?

[Wyrm]

Because the rudder provides horizontal stability, and horizontal stability is important. When only the movable part of the rudder was lost, the non-movable part was still keeping horizontal stability, so the plane was still somewhat controlable. When the whole rudder was lost, horizontal stability went south, and the aircraft became too hard to control. *tilt*

(Don't blast me too hard because I'm not a pilot by any stretch of the imagination.)

[Darth Wong]

The problem with testing and safety is that they cost money, and to be brutally honest, there is a certain dollar value that society places on human life.

[Broomstick]

The problem with testing and safety is that they cost money, and to be brutally honest, there is a certain dollar value that society places on human life.

True.

Also, the money supply is not limitless. At a certain point, paying for safety would mean you have no money left for fuel and flying would cease. Repeat for just about anything. It's something most people do not want to face, but there's inherent risk in any mode of transportation.

You can see this in railroad crossings, too - the trend in recent years has been to pile one safety device on top of another at a road/railroad crossing to increase safety. Lights, bells, whistles, lowered speed limits, better brakes... and we STILL have pedestrians step in front of trains, cars stop on the track for no good reason, etc. and get squashed. In my area they're now talking about viaducts or tunnels to separate railroad and road traffic instead of these crossings, which would be a leap forward in safety, but there's no way to convert EVERY such crossing to that - there is not enough money in the country for that. We've reached a point of diminishing returns. But even if you did that you'd still either have idiots going around the safety features and getting onto the tracks, or the occassional train or car falling off a bridge or tunnel fires or some other thing happening.

The world is not a safe place. There's no way to make it completely safe. Particularly not when you have humans wandering about the place.

[Broomstick]

Just a couple of terms, so we're all using the same language:

The vertical part of the tail of an airplane comes in two parts, one moving and one not. The moving part is the rudder. The non-moving part is the vertical stabilizer - not only is it oriented vertically, it also provides stability and control around the vertical or "yaw" axis.

It looks like in the top photo, only the moving parts of the tail fell off, while in the second the entire tail did.

Because the rudder provides horizontal stability, and horizontal stability is important. When only the movable part of the rudder was lost, the non-movable part was still keeping horizontal stability, so the plane was still somewhat controlable. When the whole rudder was lost, horizontal stability went south, and the aircraft became too hard to control. *tilt*

You're both partly correct, but the flight environment was also very important.

The first plane, an Airtransat jet, lost its rudder in cruise flight. Now, that raises some questions about the integrity of the machine (and about Airstransat, which has a history of peculiar things happening to its airplanes), but an important factor is that an airplane such as a passenger Airbus is designed to be most stable and efficient in cruise flight. There is very little or no need for rudder input during that part of a flight. The pilots were, of course, aware something was wrong in the back of the airplane, but they did not realize they had entirely lost the rudder until they were on the ground and looking at the airplane. Knowing something was amiss, they used very gentle, shallow turns (ones that could be made without use of a rudder) and very cautiously landed the airplane. There were no unusual forms of turbulence or instability in the air it was traveling through that would further complicate the situation.

Losing the entire vertical part of the tail is extremely serious - but in WWII there was at least one instance of that happening to a large bomber and the pilot still managed a controlled landing. The fuselage itself provides some stability in that axis as well. That is why the larger the airplane the relatively smaller the rudder/stabilizer is (which you may or may not have ever noticed). It is also why it's possible to design an airplane without a vertical stabilizer in the usual sense of the term.

However, the AA jet that lost its tail was on take-off - a point at which the airplane is relatively slow and climbing, and low in altitude, which cuts the safety margin much finer than in most other parts of a flight. Even so, if the air had been still they might have retained control of the airplane (a big if, but a possible one) but they were hit by wake turbulence. Airbus blamed the snap-off on the "forceful" use of the rudders by the flying pilot -- well, yes. Wake turbulence can flip an airplane over. Although a Boeing 707 has been successfully rolled (after which the test pilot was fired) no one has tried that in an Airbus, or even another model of Boeing civilian jet. No one is particulary eager to try it, espeically not with a full load of passengers who will not appreciate even momentary inverted flight. The pilot was using the force needed to keep the Airbus from going aerobatic. He was also (from the FDR) unquestionably under manuvering speed - a speed under which, by definition, you can throw the controls all the way to the stops with full force without damaging the airplane. Airbus claimed that no, you can't go full rudder one way and full rudder the other, so the pilot's training was faulty. Um... ALL pilots are trained to use as much rudder as necessary to maintain stability from early in flight training. Let's just say there is still disagreement over some of the finer points of this accident. Anyhow, the point is that not only did they lose a major flight control, they lost it in highly turbulent air, when they needed the rudder and stabilizer the most to maintain control. Fully intact airplanes have experienced complete loss of control and subsequent crash during wake turbulence encounters - an airplane missing a major control surface/stabilizer doesn't stand a chance.

The Airtransat jet lost its rudder in a mode of flight where the rudder was least critical. The AA jet lost its rudder and stabilizer in the mode of flight when they were most critical.

[Broomstick]

CHALK'S FLIGHT 101
Crashed plane's wings both had cracksInvestigators found fatigue cracks on both wings of the seaplane that crashed, killing 20 people.
BY MATTHEW I. PINZURmpinzur@MiamiHerald.com

The Chalk's Ocean Airways seaplane that crashed last month had fatigue cracks on both wings, according to a preliminary report released Monday by the National Transportation Safety Board.

Investigators almost immediately identified the cracks on the plane's right wing, which broke away when Flight 101 crashed into the ocean near South Beach on Dec. 19. The new report was the first indication of similar damage on the left wing.

''Fatigue, unfortunately, is a pretty insidious thing,'' said Grant Brophy, a private air-safety inspector and director of flight safety and security at Embry-Riddle Aeronautical University in Daytona Beach. ``This is an issue which any operator of this aircraft would need to look into, not just commercial operators.''

All 20 people were killed when the 58-year-old Grumman G-73T Turbo Mallard crashed after leaving Watson Island for Bimini. Investigators quickly focused on metal fatigue but insisted it was premature to assign blame.

A final report from the NTSB is not expected until late this year.

The preliminary report described the cracks as ``evidence of fatigue cracking in the [right wing's] lower rear wing spar cap, along the lower wing skin, and on an internal z-stringer. Additional evidence of fatigue cracking has been noted in the corresponding area on the left wing.''

The 37 Mallards still in operation were all grounded by a Dec. 30 order from the Federal Aviation Administration. Only Chalk's still flies the World War II-era planes commercially, but all owners need to provide test results showing the airframe's safety before the FAA will clear them to fly.
None have done so, according to FAA spokeswoman Laura Brown.

Chalk's voluntarily grounded its remaining fleet of four Mallards before the FAA order. Brown said the airline has talked with FAA officials about a new inspection program but has not yet submitted a written proposal.

The airline's owner, Jim Confalone, and manager, Rajan Nair, did not return calls Monday.

What this comes down to is that while the FAA thought the testing techniques and schedules were adequate for this airplane, they apparently were not. Similar fatigue cracks in the unbroken wing is a very omnious sign. You have to remember than aviation is only 100 years old, and we only have about 70 years of real-world experience with primarally metal construction airplanes. Aircraft are subjected to types of stress that other vehicles are not. We do not know everything and we are still learning.

This could also spell the end for Chalk's Airline, which has been in operation since 1919 and is the world's oldest passenger carrying service still in business. Or it was.

In case you were wondering, this is what an intact Mallard looks like. In this case, the airplane is owned by an Australian company. Yes, it's supposed to be in the water - it's primarially a seaplane with some amphibious capability.

[Red Star]

Metal fatigue cracks in compression are much worse, because you can't even see them upon close inspection. In one case an oil rig went down because one of its support members failed due to fatigue despite having been recently inspected for fatigue cracks with ultrasound equipment.

If there is a crack, an ultrasound will pick it up better than a x-ray or penetrating dies. The NDT (thats what their called at the place I work) guys must have missed the crack completely.
I don't know how good any inspection equipment would be for compression other than maybe irregularities in the pattern that the ultrasound would show on its screen.

[Broomstick]

NDT = Non Destructive Testing

I'm not sure what sort of testing they were using on the Mallards, but it missed the cracks on both sides of the wings. Maybe there were issues with getting the testing equipment inside the wing. As things stand now, there's talk of pulling the wings off all of them and peeling the skin off to really get inside. Which will be hideously expensive, of course.

There's an argument to be made to simply retire all the Mallards, but of course, the owners won't be happy about that at all. And one of the reasons that Chalk's was using these airplanes was because there really haven't been any passenger-carrying seaplanes built since the 1940's. It's a very limited niche.