Asiana 214 Survival Factors

  • E-Mail this Article
  • View Printable Article
  • Text size:

    • A
    • A
    • A

Standards must be slipping at CNN. Early Saturday morning, they didn’t have the United 777 diversion to Midway story. I expected the network to have summoned their stadium full of talking heads to connect that incident to--what else?--Malaysian 370. I figured they could spin a couple of days of 24/7 speculation out of the Midway incident…you know, smoke in the cockpit, power failing. But not so far. Maybe I shouldn’t give them any ideas. We’ll see what develops.

Meanwhile, my attention was focused on another 777, the Asiana Airlines flight that crashed in San Francisco last July. To no one’s particular surprise, at least in the aviation community, the crew took the hit for mismanaging the airplane's vertical profile during the approach and for ignoring or not having a good understanding of the automation, especially the autothrottle modes. The full report will shed detail on other factors and it’s due for publication later this month.

My longtime friend Bill English was the NTSB investigator in charge on Asiana 214 and he stopped by the other day. We both agreed that this accident was really a GA-like accident. Break it down and it’s not much different than a Cirrus being flown into terrain on autopilot or a Cessna 210 driver getting low on the PAPI, cratering the airspeed and mushing it in. I suspect something will be made about this being another “automation accident” and while there’s some merit to that, it’s really all about consistent, basic airmanship, or lack thereof. Three qualified and trained pilots saw the PAPI going red, saw the airspeed decaying and simply rode through it. As such, I find that aspect of the accident … uninteresting. It’s a human factors fail. What else is new?

What’s more interesting, at least to me, are the survival factors related to the airplane itself. This is only the second in-flight hull loss for the 777, of about 1200 flying since the airplane entered service in 1995, almost 20 years ago. The first hull loss, recall, was at London’s Heathrow in 2008, when both engines reverted to idle power because of ice crystals in the fuel system. There were injuries, but no fatalities and no fire, despite significant damage to the aircraft. An EgyptAir Triple Seven burned at the gate in 2011 and was written off and, of course, there’s MH370, which seems less and less likely to be considered an accident. It’s unlikely to mar the type’s insurance record.

Because all but three people survived the Asiana 214 crash, it’s quite natural to think of it as just a skid down the runway, sans gear and engines. But that it definitely was not. Bill told me that when he got the go call, he was told only that a mass casualty event had occurred at SFO and an airplane had cartwheeled. He hadn’t seen any news coverage and hadn’t seen the video.

“Then I saw the photo from above. It looked like an airplane…it was canted off one side of the runway. How can it be together like that? So we were not believing the cartwheel story. In fact, I don’t think I saw the video until later the next day,” Bill told me.

The video in question came from a security camera. It yielded footage reminiscent of the United 232 crash in Sioux City, Iowa, in 1989, which also looked like a cartwheel, or close. But subsequent investigation of 214 revealed what was more likely a turning pirouette of sorts. The left wing tip scrapped, but the radome never touched pavement. Bill thinks the fact that the tail departed while the engines were spooling up in a last ditch effort to save the approach may have accounted for that. It probably had a significant upward moment at the point of impact and with the tail gone, the CG shifted rapidly forward.

Whatever the reason, the impact forces were enormous, well exceeding the 16Gs the cabin floor rails are rated to carry.  “We can certainly see that the airplane survived the impact and the slide down the runway pretty much intact. The center section was completely intact, the forward section had essentially no damage. In the aft part of the airplane, the floorboards gave way, but you could certainly see why because of the enormous forces. But there was survivable space,” Bill says. The impact was violent enough to produce head injuries to people striking either adjacent seats or seat arms across the aisle

Two of the three passengers who didn’t survive appeared not to have been belted in and were ejected. It’s speculation to say they would have survived if they had been belted, but it’s hardly unreasonable speculation. One passenger was struck by a door that came off its hinges during the impact sequence and likely would not have survived. Of 307 aboard, 182 were injured, 12 critically.

Post-crash photos revealed that the hull burned, but it wasn’t a fuel-fed fire. The right engine got trapped under the wreckage and a punctured oil tank leaking onto hot components appears to have started a small fire that transferred heat into the cabin. But the fire itself didn’t penetrate the cabin and wasn’t a survival or evacuation factor. Neither were the evacuation slides, two of which inflated inside the cabin. Most passengers self-exited, but some had to be helped by rescue personnel. Despite the slowly advancing fire, they had sufficient time to rescue everyone. The failed slides had been subjected to g-forces far beyond their design loads.

The investigation revealed a powerful lateral movement to the left during the impact sequence and a high number of thoracic spinal injuries. Although the mechanism isn’t understood or at least documented, it sounds likely that passengers were simply wrenched violently to the left in their seats. Subsequent investigation may reveal more about this and perhaps inform how seats and seatbelts can be improved. Would shoulder harnesses have helped?

In the meantime, it’s fair to conclude that Boeing applied what it learned from building airplanes and investigating crashes to make a safer, stronger airliner. Crash and impact forces that were barely understood during the 1970s and 1980s were yielding to the computer-aided design used to build the 777. It hung together through a horrific impact sequence. And it hung together after that, too.

“Weeks after we had done the investigation on the airplane, the wreckage had to be moved off the airport. There’s not that much room at San Francisco,” Bill told me.  “So we had to break it down into smaller pieces. It was an enormous effort. It’s not easy to break apart a Triple Seven. It looked like something out of the movie Transformers to get that center wing box apart.”

After I talked to Bill, I got wondering whether the Triple Seven has the best accident record of any transport category aircraft. It doesn’t quite, but close. According to Boeing’s own numbers (PDF) the Airbus A380, the 747-8 and the 787 have better safety records than the 777 among wide bodies, but none have close to the fleet experience, having accumulated fewer than a million departures, at least through 2012. Surprisingly, despite flying in higher risk operations—more weather, more takeoffs and landings—the CRJ-series regional jets and Boeing's own 737-800-900 series have a little better numbers than the Triple. So does the 717, but only about 150 of those are flying. The Airbus 320-class is comparable to the 777, albeit with many more hull losses.

But it’s a quibble hardly worth the pixels to describe. Modern airliners, wide bodies like the 777 included, have become exceptionally reliable and safe thanks to lessons written in flesh and blood during the 1960s and 1970s. It’s hard to imagine them improving much, but they’re very likely to do just that.

So much so, in fact, that if Bill retires from the NTSB in five years, which he might do, Asiana 214 could easily be his last major U.S. accident investigation. That notion would have been unthinkable 30 years ago. But it isn’t now.

Bill will be AirVenture with a presentation on Asiana 214 survival aspects. See him in the Federal Building on Saturday, August 2 at 11:45 a.m. and catch him on EAA Radio on Thursday, July 31 at 12:15 p.m.

Join the conversation.
Read others' comments and add your own.

Comments (29)

I honestly believe that Boeing has had a good run of structurally strong and safe aircraft designs. However, the integration of glass panels and Digitized Automation systems and their torrent of information affect the safety standard by their extent and complexity.

The new generation of flight crews are perceived by some as program or system managers rather than skilled "stick-and -rudder" aviators, corroborating the argument that hand flying skills are degrading - mainly due to autopilot dependency and insufficient training.

During the last 12 years several examples surface, the Air France AF447, SFO Asiana flight and arguably American Airlines AA587, all reported inept PF and flight crew operations as contributing factors. The other factors were blamed on incomplete company training, structural design faults (A300 series vert. stab) and ergonomic deficiencies.

So I ask, how many inept crews are flying safe aircraft and who has a safer design, Boeing or Airbus?

Posted by: Rafael Sierra | July 13, 2014 7:10 PM    Report this comment

" We both agreed that this accident was really a GA accident."

Hmmm.... an airline pilot screws up on the job and it's a "GA accident". No wonder their safety record is so good!

Posted by: A Richie | July 14, 2014 8:46 AM    Report this comment

Wow, to ask which one is safer - Boeing or Airbus - could be a whole separate discussion. Just as you have people that prefer Boeing airplanes over Airbus (just because they like one and love to hate the other), and vice versa, I am sure that there are pilots that prefer one manufacturer's design philosophy over another. I would bet you can find a whole bunch of "experts" to say Airbus has a better automation design, the same number favoring Boeing. I think it comes down to this - the pilot's knowledge of the system and flying the aircraft within the limitations. Both manufacturers make a safe airplanes but the pilot has to know the systems, the limitations, and needs to have his/her head in the game - meaning even when the hands are not on the controls, you have to be "flying" the airplane as if they are. Before Asiana 214 descended below the glide path, they should have noticed an issue and started a correction. Before they went below the approach speed they should have realized there was an issue and made a correction. These pilots were passengers because their head was not in the game...they were not flying the airplane.

Posted by: Mark Hancock | July 14, 2014 11:01 AM    Report this comment

There is no one better than Bill English to evaluate which one is safer. Perhaps Bill can enlighten us after he retires. Paul B. states "The Airbus 320 class is comparable to the 777, albeit with many more hull losses." suggesting Boeing superiority. From my perspective an example of a rushed into service design is the A300 series (AA587) where the Vertical Stabilizer structural strength limits were exceeded allowing for an inflight break ending in a "mass casualty event".

Airbus (read AA587) denial and influence peddling accusations are still in the dark side. My impression then was that Airbus was not as structurally safe as Boeing but with the introduction of the B777 and B787 and its thermoplastic content (similar to Airbus) and fly-by-wire systems, I wonder what the future holds for Boeing's safety record.

Posted by: Rafael Sierra | July 14, 2014 2:38 PM    Report this comment

"Airbus (read AA587) denial and influence peddling accusations are still in the dark side. My impression then was that Airbus was not as structurally safe as Boeing but with the introduction of the B777 and B787 and its thermoplastic content (similar to Airbus) and fly-by-wire systems, I wonder what the future holds for Boeing's safety record."

My understanding of AA587 was that the composite structure of the vertical stab broke beyond its design limit loads, and that the problem came about because the FBW system allowed for full stop-to-stop deflection of the rudder, even at the 200-something knots the aircraft was flying at.

One thing to note is that there is a difference between composite structures and thermoplastics, though I've noticed many people erroneously refer to the two as the same thing. There are also many ways to construct composite structures, so even generalizing them together incorrectly describes these structures.

Posted by: Gary Baluha | July 14, 2014 3:36 PM    Report this comment

"Compared with traditional metals, plastics and thermoset composites, fiber reinforced thermoplastic are lighter, tougher, stiffer, and more sustainable and can be produced in high volumes-- with the right process.

The Fiberforge process converts pre-impregnated thermoplastic tapes into advanced composite structural parts that meet the demands of almost any high performance application, ranging from aircraft parts to backpack frames.

Thermoplastic advanced composites compared to other materials: 60% lighter and 600% stiffer than steel 30% lighter than aluminum 200% tougher than thermoset composites 500% stiffer than injection molded plastics 60% less scrap during production than sheet "

Gary Baluha, I spent many years designing thermoplastic and thermoset product for commercial and military apps.

Posted by: Rafael Sierra | July 14, 2014 3:46 PM    Report this comment

Well, it sounds like great pilots in a great airplane got caught in an unfortunate situation of actually haviing to fly it. At least they got part of the aircraft to the runway on the first impact.

Posted by: Tod Dickey | July 15, 2014 1:48 AM    Report this comment

"My understanding of AA587 was that the composite structure of the vertical stab broke beyond its design limit loads, and that the problem came about because the FBW system allowed for full stop-to-stop deflection of the rudder, even at the 200-something knots the aircraft was flying at."

And that should tell you a lot about the shameful lack of sophistication of Airbus' FBW software.

Posted by: Thomas Yarsley | July 15, 2014 7:40 AM    Report this comment

Rafael, The point I was trying to make was simply that "thermoplastic" isn't the same as "thermoplastic composite", and "composite" in general is just a term that describes mixing different materials together (simplified meaning). Didn't mean to imply you didn't know what you were talking about.

Posted by: Gary Baluha | July 15, 2014 8:11 AM    Report this comment

Thank you Gary, I thought you were joining ranks with my wife for a moment. The question still remains who has the most safe aircraft, Airbus or Boeing? Airbus is not my favorite but now Boeing is getting their share of problems with the 787 and since they have more plastic than Airbus I am not comfortable in them as I am quite aware of thermoplastic delamination, disbonding and aging. The 777 had initial problems but the wiring and battery problems continue. In the meantime Airbus benefits by Boeing's pains. So I ask, who has the safest structural, electronic, and ergonomic design?

Posted by: Rafael Sierra | July 15, 2014 9:12 AM    Report this comment

"So I ask, who has the safest structural, electronic, and ergonomic design?"

What do the numbers tell you?

Posted by: Paul Bertorelli | July 15, 2014 10:04 AM    Report this comment

There's never been an in-flight breakup of Diamond -- all-composite glass fiber reinforced plastic (GFRP) and carbon fiber reinforced plastic (CFRP) -- DA20/DA40/DA42 aircraft. The same can't be said for aluminum Piper or Cessna or Beech GA aircraft.

I'll gladly ride in one of the newer composite airliners.

Posted by: Dave Passmore | July 15, 2014 11:01 AM    Report this comment

Paul, the numbers tell me that Airbus has a problem, from the A300 to tge A380, and that's why I don't do Airbus. The design criteria Bill English referred to during the AA587 investigation was that from the DC3 era. The vertical stabalizer assembly load limits were set for a much shorter aircraft.

Posted by: Rafael Sierra | July 15, 2014 1:25 PM    Report this comment

They should never have quit making airliners from tube and fabric.

Posted by: Richard Montague | July 15, 2014 1:48 PM    Report this comment

Dave, Diamond DA40 is the only composite certified airplane that does not have an airframe life limit. A Cirrus, after 12,000 hours is not airworthy. Airbus and other plastic aircraft should have an airframe time limit as well.

Posted by: Rafael Sierra | July 15, 2014 2:38 PM    Report this comment

"Paul, the numbers tell me that Airbus has a problem, from the A300 to tge A380, and that's why I don't do Airbus."

What numbers?

Posted by: Paul Bertorelli | July 15, 2014 3:13 PM    Report this comment

Raf:

This crap about airframe life limits is just that - crap. It has NOTHING to do with structural strength or the expectation of actual duration of attainment of standards. It's simply a method of certification, and one that's considerably easier to do than the alternative "fail-safe" methodology.

I've owned a Piper Tomahawk for 27 years (and flew them for many years prior to that). Its wing has a main spar that's a tip-to-tip I-beam - you have to see it to believe it. And yet, it's life-limited. Why? That's right - Piper took the easier certification route, to get the plane to market faster and cheaper. The idea at the time was that nobody would care that, sooner or later, some Tomahawks would achieve 11,000 hours of flight time. Honestly, most original owners of Cirrus SR-2x airplanes probably would consider them to be disposable by 12,000 hours, regardless of how many years it took them to achieve that usage - because anybody who can afford three-quarters of a million dollars for an unpressurized, fixed-gear single, likely can afford (or simply wants) something new by then.

Posted by: Thomas Yarsley | July 15, 2014 3:46 PM    Report this comment

Thomas: In plastic aircraft the aging of thermoplastic materials affect structural strength. I believe that airframe time limits are reasonable measurable factors just as strongly as you consider it "crap" or nonsense. Aging, delamination, disbonding and brittleness are real reactions negatively affecting the original strength of the thermoplastic composites. It would be irresponsible to disregard the deterioration caused by environmental temperature differentials, vibration, application and quality control of materials and workmanship. Inspection of composite structures are more difficult than in metal, much deterioration (delamination, disbonding) may go undetected - this has happened and it is on record. Air Transat Flight 961 is a good example. AA587 fatalities (Total: 265) including 5 on the ground that could have cared less about our "passion" are ashes now. Aircraft plastic fuselage structures must be unquestionably safe.

The Cirrus example I cited was to bring out the safety concerns regulating authorities have of new technologies and applications - when in doubt establish safeguards. I welcome them, safety standards must be practical anticipating incidents or fatalities.

Posted by: Rafael Sierra | July 15, 2014 7:27 PM    Report this comment

Thomas: In plastic aircraft the aging of thermoplastic materials affect structural strength. I believe that airframe time limits are reasonable measurable factors just as strongly as you consider it "crap" or nonsense. Aging, delamination, disbonding and brittleness are real reactions negatively affecting the original strength of the thermoplastic composites. It would be irresponsible to disregard the deterioration caused by environmental temperature differentials, vibration, application and quality control of materials and workmanship. Inspection of composite structures are more difficult than in metal, much deterioration (delamination, disbonding) may go undetected - this has happened and it is on record. Air Transat Flight 961 is a good example. AA587 fatalities (Total: 265) including 5 on the ground that could have cared less about our "passion" are ashes now. Aircraft plastic fuselage structures must be unquestionably safe.

The Cirrus example I cited was to bring out the safety concerns regulating authorities have of new technologies and applications - when in doubt establish safeguards. I welcome them, safety standards must be practical anticipating incidents or fatalities.

Posted by: Rafael Sierra | July 15, 2014 7:27 PM    Report this comment

Raf:

I'm very aware of the aging concerns regarding "plastic" materials. And I take no issue with including life limits in the arsenal of certification tools.

My point was directed to the commonly-held (and frankly, the too-often-proselytized) belief that components that are certificated under safe-life rules necessarily are inferior to components certificated under fail-safe rules. They are not. That is the "crap" to which I referred.

In the case of AA587, it didn't matter whether the tail was made of aluminum, plastic, or unobtanium. Airbus' crappy FBW software failed to protect the airframe from structural failure.

Safe-life limits may make people feel better, but they do nothing to protect against the dangers of undetected structural compromises. Regardless of certification methods, it is incumbent upon the industry to be vigilant and to develop reliable means of non-destructive field testing, to ensure continued performance of composite structures.

Posted by: Thomas Yarsley | July 15, 2014 8:29 PM    Report this comment

Thomas, the AA587 A300 was not fly-by-wire. No rudder deflection limits. The pilot flying did not exceed Va while using rudder - this subsequently changed training criteria. Airbus came out with an advisory suggesting not using rudder during roll outs as a remedy. The design of the vertical stabilizer assembly is the same in all Airbus aircraft including the A380. Not a good thing.

Posted by: Rafael Sierra | July 15, 2014 9:19 PM    Report this comment

Mire about thermoplastic composites.

"Most experts agree that composites are stiff (non-ductile) when compared to metals and have differing thermal coefficients than metals at attachment points. Also, air and water are invariably entrained during the manufacturing process. As the aircraft operates over a wide range of external temperatures and pressure altitudes, these water and air deposits expand and contract, eventually causing delamination (fatigue) within the composite. These, plus other known and unknown factors affect composite material to one extent or another, ultimately reducing its utility.

According to Dr. Debra Chung, Director of the Composites Research Lab at the University of New York at Buffalo, composites thought to have a useful life of 30 years may, in fact, only last 1/3 of that original estimate because they become brittle and lose elasticity.

So an aircraft which is expected to operate for 30,000 cycles, may only last 10,000. Charles F. Marschner [1], a pioneer in the use of plastics in aircraft structures, reminds us that "the laws of physics and mechanics are immutable - today's composites still have no ductility as do wrought metals.""

Posted by: Rafael Sierra | July 16, 2014 4:06 AM    Report this comment

And yes, the 777 has an outstanding safety record.

Posted by: Rafael Sierra | July 16, 2014 4:18 AM    Report this comment

The point I was trying to make Raf, was that you're saying Airbus has a problem, but if you examine Boeing's own safety data, the 318-320 series has a comparable safety record to Boeing products. If there's a safety problem, the data doesn't support it. The older 300, 310 models have higher rates, but then so do Boeing models.

Also, we should point out that composite failure was not given as causal or contributory in AA587.

Posted by: Paul Bertorelli | July 16, 2014 6:33 AM    Report this comment

Raf:

I should not have used the term "FBW" with regard to the AA587 vehicle. More properly, the A300 has "boosted controls," which have been in use for a very long time. But any use of hydraulics, pneumatics, electrical, etc., that delivers more gain to control-surface movement than would be afforded by a direct cable / pushrod link, must be attended with some sort of gain control that precludes over-stressing the airframe with simple movement of the flight-control surfaces.

There are lots of ways to do this, with varying degrees of sophistication and complexity. But it must be done. Airbus failed to do that - its control system allowed a pilot to separate the vertical fin from the fuselage - using nothing more than his two feet. Personally, I would never design any control system that would permit that to happen. Airbus did.

With regard to the "stiffness" of composite structures, one must consider each complete assembly, rather than the flextural modulus of the primary composite matrix. The B-787 wing is a good example of a very flexible assembly that's built from a (composite) material that not many would characterize as "very flexible."

Given the tight SPC environment in which Boeing's composites are laid up, I'm less concerned about entrained gasses and vapor-phase liquids than you appear to be. Resin sublimation also doesn't concern me much, nor does delamination where filament-winding lay-up is employed. Resin crystallization is another issue.

I wasn't involved in Boeing's design, but if I were, I would have included "coupon" areas throughout the structure, to provide future ability to conduct destructive testing and evaluation throughout the service lifetimes of the vehicles. There's precedent for this from long ago - surplus "access panel" perimeters included in areas of fabric-covered aircraft. The material that spanned them could be excised for examination and destructive testing, and the hole could be filled with a stock cover. It helps to know when your linen is about to turn to dust!

Filament-wound high-pressure air tanks have been around for quite a while now. FEA now is ubiquitous, although its predictions are only as good as the fidelity of its underlying materials-modeling. Ultimately, safety is about failure modes. Composite layup construction seems to compare favorably with traditional riveted aluminum. Friction-stir-welding of thin aluminum (a la Eclipse) still scares me more than does adhesive bonding of similar structures. But I have a lot more experience with structural adhesives, so that probably accounts for my higher levels of confidence.

Posted by: Thomas Yarsley | July 16, 2014 10:30 AM    Report this comment

Paul, I understand as I have reviewed the Boeing vs Airbus safety record. What I really want to know is who has a safer design, Boeing or Airbus? Metal vs composites? Is Boeing going safer with the 787?

Perhaps you can explain this to me.

Posted by: Rafael Sierra | July 16, 2014 10:38 AM    Report this comment

I'm quite certain I can't explain it.

For many years, in covering the TB20/TB21, we have bitched about the gullwing doors being a bad safety design because in the event of a turnover, you can't get them open. And for just as many years, this has never been a factor in any accidents. So perhaps it's a question of not-the-ideal design having no bearing on safety and accidents. This could be true of metal versus composites.

Personally, I have a strong Boeing bias because I like the way they've designed their FBW control laws. But at some point, reality must intrude and when you look at the safety/accident record, they are often comparable. To me, that's the only fair and accurate way of looking at such things.

Fortunately, the airlines collect and process very accurate data so statistical models are not just accurate, but also accurately predictive in some cases. We can only dream about this in GA.

Posted by: Paul Bertorelli | July 16, 2014 4:40 PM    Report this comment

On second thought, put this in the GA context. Diamond's DA40 has the best single-engine safety record by far. It's a composite airplane. But would it be just as safe if it were all metal? In other words, is the design features or the material that makes it so safe?

I think it's a combination.

Posted by: Paul Bertorelli | July 17, 2014 9:32 AM    Report this comment

Looks to me like the comments diverted a lot from the blog statements. The bottom line of the Asiana accident is that the pilots stopped (or never really started) flying the airplane. Instead they poorly monitored the automation that was flying the airplane and didn't react promptly when that automation wasn't doing its job properly. The lesson, it seems to me, is that if you're going to allow the automation to do the flying, you better be monitoring it accurately and be ready to hit the "off" button and fly it manually when things don't look right. To that extent, whether it's a 121 flight or a 91 flight, the pilot's job is much the same.

But humans being what they are, when someone or something is doing the work, it's way too easy to sit and stare and not be able to react quickly enough. It's a little like a flight instructor who isn't quite ready to take the controls when the student botches a landing, although he/she could have prevented the subsequent runway excursion had he/she reacted more quickly.

That the airplane was built well certainly contributed greatly to the minimal loss of life and injuries. But just as great a contribution was that although the pilots terribly botched their jobs, the airplane hit the runway where it did and not a whole lot sooner. There were many fortuitous pieces to this accident, but whether the airplane was metal or plastic or some combination probably is mostly immaterial.

Posted by: Cary Alburn | July 18, 2014 12:53 PM    Report this comment

Add your comments

Log In

You must be logged in to comment

Forgot password?

Register

Enter your information below to begin your FREE registration