Intentionally crashing a Boeing – what did we learn?

On December 1, 1984 a remotely piloted Boeing 720, loaded with specially formulated anti-misting Jet A, was intentionally crashed at Edwards Air Force Base to determine if the fuel would preclude or suppress a post crash fire long enough for occupants to escape. It was a bold but ill-conceived experiment that went up in smoke.

Dummies in Boeing 720
“Ladies and gentlemen, please buckle your seat belts…”

In addition to the anti-misting kerosene (AMK) evaluation, the controlled crash also provided data on how passenger seats and other structures performed in such situations. Instrumented dummies were seated in the cabin to assess acceleration forces and cameras documented fire propagation and how well other fixtures held up. It was well planned and carefully rehearsed over four years including multiple remotely-piloted approaches to 150 feet above the ground, 16 of which included engines running on anti-misting kerosene.

Engines had to be modified with degraders to chop up the AMK’s long molecules so fuel would flow reliably into combustion chambers and burn like regular Jet A. Proving flights were a cautious, step by step process, incrementally feeding the AMK from a few tanks to a few engines to be sure engines ran properly.

Airlines were deeply skeptical about the whole idea and very concerned about its costs and practicality. Going forward with such a program meant, at the very least, adding more steps to fuel refining and costly fleet-wide fuel system retrofitting to accommodate the AMK’s long fuel molecules. All of this to address those extremely rare events where suppressing or delaying a post crash fire would allow passengers to escape in an otherwise survivable accident.

The industry view was that the money could be better spent on accident prevention rather than adding costly mechanical complexity to prevent what might possibly happen in rare post crash events. Instead airlines advocated better automation, cockpit displays and warning systems as a better use for the money. More on this later.

NASA worked diligently on the project, methodically fixing remote control bugs and refining control techniques to where they were confident the old Boeing could be flown wings level into eight fixed barriers designed to slice open fuel tanks but leave the fuselage intact.

Boeing 720 slapdown
This could hurt…

Finally, with all details complete, the crash date was set. Word went out to the airlines, manufacturers and other interested industry groups to come see the fruits of NASA’s efforts. And so, representing Pan Am, I gathered with others at Edwards Air Force Base on that cool December morning several miles from the Rogers Dry Lake runway where NASA ’s remotely controlled Boeing 720 loaded with 76,000 pounds of anti-misting kerosene would end its last flight.

The plane lifted off, retracted its landing gear, climbed to 2300 feet then banked around and lined up to land wheels up on the spiked runway. We watched through binoculars as the 720 began its descent on a slightly steeper than normal 3.8 degree descent toward the runway. Standing beside me was Alex Ogston, an old timer British chemical engineer who worked for Standard Oil in World War II helping develop the 100 octane gasoline that contributed to the Spitfire’s success in besting the Germans in the Battle of Britain.

As we waited there, Ogston chatted about those long ago times and related how Messerschmitt 109s had to make do with 87 octane gas while the Brits’ 100 octane fuel allowed higher manifold pressure and more power for their Merlin engines, giving them a narrow edge over their adversaries. Ogston was incredulous about what he said was “NASA’s silly effort to keep jet fuel from burning.” I just nodded and said: “Well, we’ll know shortly.”

Boeing crash
Test failed?

Nearing touchdown we noticed the Boeing bank left and right then strike the ground slightly left wing down. From our distant vantage point we couldn’t see precisely what followed other than almost immediately a monstrous fireball erupted as the plane slid along. The old engineer standing beside me was right. Liberating tons of jet fuel in the presence of an ignition source will result in a large fire ball. Fire fighting vehicles arriving on the scene were no match for the conflagration and the plane burned over an hour in spite of their efforts.

The 720’s wing wobbling dutch roll (common in swept wing aircraft) was at the root of the pilot’s control problems. Seeing that a wings level touchdown was doubtful, the remote pilot spooled up the engines apparently trying to go around but couldn’t complete the maneuver in time. The plane struck the ground left wing down in a left skid at full thrust instead of being at idle for landing. It then slid into the barriers, one of which sliced through the number 3 engine and passenger cabin, providing a flame path into the fuselage. The botched experiment highlighted the fallacy of carefully engineering a crash scenario to serve as the basis for retooling airliner fuel systems and reformulating jet fuel specifications.

The fireball and post crash analysis dramatically confirmed industry skepticism about AMK, and pointed up its shortcomings as a viable safety enhancement in the real world. The effort was abandoned. But it wasn’t all for naught. Analysis of fire propagation in the cabin led to new standards for fire blocking materials in passenger seats and highlighted the need for faster flight recorder data sampling rates.

The FAA estimated that 25 to 28 of the cabin’s 113 occupants might have been able to exit the cabin before dense black smoke completely obscured visibility. Escape time varied from five seconds in the forward cabin to 20 seconds further back. I’m skeptical of the FAA’s survivability estimates in such a fiery accident scenario because passengers often wear clothing and footwear providing almost no bodily protection and some seem only marginally able to maneuver into and out of seats even in normal circumstances.

In contrast to that long ago AMK experiment, consider how today’s well engineered terrain awareness warning systems (TAWS), also known as enhanced ground proximity warning systems or EGPWS, have largely prevented the kind of accidents AMK was intended to make survivable.

Enhanced ground prox systems were first installed in air carrier jets in 1997 and are now in over 55,000 airliners, corporate jets, turboprops, helicopters, business aircraft and military transport aircraft around the world. FARs mandate them for passenger airplanes operating under FAR Part 121, turbine powered planes operating under Part 135 with 10 or more passenger seats and Part 91 turbine powered planes with six or more passenger seats. TAWS installations can include a worldwide terrain and obstruction database and cover all airports with paved runways 2200 feet and longer, although some systems are less inclusive depending on user needs.

Garmin terrain alert
These warnings did more for safety than the fanciest fuel projects.

TAWS warns pilots of terrain and obstructions with visual and audio alerts plus, on Part 121 and Part 135 aircraft, color-coded situational awareness terrain displays. TAWS also warns of flight dangerously close to terrain, excessive bank angle, excessive deviations from the ILS glideslope or excessive deviations from the approach descent path – as well as descents after takeoff. Since EGPWS was introduced 20 years ago, the airline hull loss rate for Western built airliners has decreased about 2.5 times.

What TAWS can’t do is convince overly headstrong pilots to heed warnings.

For example, on May 9, 2012, a brand new SU95-100 equipped with TAWS flew into a mountainside during a demonstration flight in IMC while the pilot in command listened to “terrain” warnings and finally, “pull-up” warnings for 36 seconds before impact.

To better understand where and under what circumstances significant airliner flightpath deviations occur, Honeywell analyzed five years of escape activations (2011 through 2015) on glass cockpit airliners equipped with their TAWS. There were 224 final approach premature descent events extracted from about 24.38 million flight legs operated around the world. None were reported by pilots and air traffic controllers. The event data covered the period 20 seconds before the alert through 10 seconds after and were de-identified so that they only could be used for safety analysis.

How many of these premature descents would have ended in an undershoot accident is impossible to know but it’s comforting to know TAWS is doing its job around the world by alerting pilots in a manner that results in a successful avoidance maneuver. And because thoughtful regulators acknowledged the impracticality of AMK as a safety enhancer, we’re not saddled with an unnecessary, unworkable fuel additive which would neither prevent accidents nor materially increase post-accident survivability.

9 Comments

  • One of the problems facing scientists in the modern era is pressure for each and every experiment to turn out with ground breaking results. In the real world, you need to get a lot of “no” before you reach “yes”. Otherwise, researchers will only go for very cautious experiments and do questionable things to force outcomes. Ease up on poo-pooing experiments with negative results. I know they are expensive, but they all add to the body of knowledge.

  • Rebecca is correct. As a research engineer early in my career I was getting very frustrated and downhearted about the failures on a particularly challenging project until my senior engineer gently reminded me that there is no such thing as a failure, we learn more from our failures than from our successes, and everything adds to the overall body of knowledge. Although in this case the mission objective was AMK, I’m quite certain that a considerable amount of knowledge was gained in many other areas, including cabin safety. Whilst this experiment might well have been expensive, as someone once remarked, if you think safety is expensive, have you priced lawsuits recently?

  • Agreed with Rebecca and Adrian. The intent, with AMK, of developing a safer aircraft fuel is admirable, given that most of the deaths from aviation accidents not due to trauma are due to fire and smoke, was very good. It didn’t work. It was worth the crashing of an obsolete jet aircraft that otherwise would have ended up in a boneyard or parted out.

    It’s still worthwhile looking for alternative fuels that don’t readily roast or asphyxiate humans on crashed aircraft.

    Fire survival is a key benefit of using hydrogen fuel cells in both aircraft or land vehicles – being lighter than air, any hydrogen that escapes a fuel tank breach simply dissipates vertically, instead of hanging around to roast human beings. Ditto with liquified natural gas, which also is lighter than air and will quickly dissipate in the event of a fuel tank breach.

  • Thanks guys. Also, as a flight instructor, I do agree with the author that you get a lot more safety “bang for your buck” when you focus on crash prevention as opposed to crash survival. I also agree that developing jet fuel that won’t burn was ill-conceived (although the concept of highly evaporative fuel is tantalizing!) That being said, Adrian and Duane are right that this experiment was probably worth doing.

  • So far the comments posted have been negative (or maybe better said as “less than positive”) of Arnold Reiner’s article. I am writing in support of Mr. Reiner. First, let us not loose sight his articles are written from an opinion basis rather than that of a reporter. Like you, he certainly has a right to his opinion, as such, this article is valid. I also agree with the underlying message Mr. Reiner expresses – that of government’s tendency to tunnel-mindedness. Yes, it is entirely possible all the effort put into the remote piloting of the Boeing 720 is today realized in both civilian and military remotely piloted vehicles. But look at the initial premise of the whole experiment – produce a combustible fuel that will not burn. Ponder that, then tell me that isn’t the very definition of the phrase – “I’m from the government and here to help you.”

    • Gregg – there were no comments here that berated the author for his remarks. The author did in fact “poo-poo” the experiment itself .. that was the entire premise of his post. The negativism here was completely on the post author’s part, not the commentors. We simply have a different perspective, and made no personal attacks on the author.

      Secondly, you misconstrue the intent of the AMK experiment, and of the worthy goal of examining other fuel types in an effort to reduce fatalities from post-crash fires. It was not an attempt to prove the impossible – that a combustible fuel might not combust. You are practicing what is called “straw man argument”. Set up a silly straw man and prove your intelligence by easily knocking him down, when the straw man is not actually relevant to the discussion.

      No – the intent of the AMK experiment was to demonstrate whether a combustible fuel can be made such that is retards unintended aircraft post-crash fires, or not. Perhaps the experiment could have been conducted more efficiently on a laboratory scale and produced the same results as it did full scale, and consume less resources as a result. But the intent of the experiment was not silly as the author and you suggest it was … and it was not the straw man that you constructed, i.e., a combustible fuel that will not combust.

      There are in fact fuels that are less likely to both concentrate in the “people compartment” and combust rapidly, and thus quickly roast the occupants in an uncontrolled structural fire. Diesel fuel, for instance, is less prone to out of control vehicle fires than is gasoline. That is one reason why diesel engines are often preferred for safety reasons over gas engines in boats. Aviation gasoline is one of the most combustible hydrocarbon fuels made because it is so volatile.

      As I pointed out in my comment, a more fruitful area of safety R&D that assuredly will reduce the destruction of post-crash fires is any kind of fuel that is lighter than air. Like hydrogen (as used in fuel cells), or liquified natural gas. Both of which will naturally dissipate vertically when released in a vehicle or aircraft crash.

      There may well come a day when such fuels predominate … and people decades from now will scratch their heads and wonder why we 20th century and early 21st century fools would risk our lives with such dangerous fuels as gasoline and JP4, with the very real risk of being roasted alive en masse during even minor accidents, as has repeatedly happened in aviation over the decades. And yes, these alternative fuels are still quite combustible.

  • I think the point of the story is not to criticize the effort of this experiment, but to point out how technology really works. It’s often the slow march of basic technology that one day grows up (EGPWS and TAWS systems) that start making a difference long before grand experiments pay off. It’s not a criticism, just an interesting historical example – we can’t always plan safety improvements as carefully as we’d like to.

  • Just to clarify, I didn’t say the experiment was “silly, ” it was the fuel engineer standing beside me who expressed that view. My misgiving and the opinion of many observers, assuming the test succeeded, was the cost vs. benefit compared to spending finite dollars on preventive safety initiatives like those discussed in the article.

  • I was in Seattle FAA Certification when this test was conducted. Everyone except me said the test was a failure. I said it was a success. In response to the strange looks I got my reasoning was simple.

    Had the Boeing 720 NOT burst into flames the all-too-obvious conclusion that most likely would have been drawn was that AMK worked! Following this “successful” result Big Brother would have begun to require the more costly production of AMK and the airlines to modify engines to operate reliably with jet fuel erroneously believed to possess “Anti-Misting” qualities.

    That could have been the tragic consequence. We’d have all been duped into flying around with fuel we believed was NOT going to burst in flames in passenger airliners with REAL people instead of crash dummies. And the truth about AMK would not have been discovered until an airliner burning AMK instead of JET-A crashed and burst into a fireball.

    That said I think the author and all those who commented completely missed the point which was that the test was successful in proving that AMK was a failure.

Leave a Reply

Your email address will not be published. Required fields are marked *