Siga-me

Roberto Henry Ebelt

The article below was published in THE ECONOMIST on April 15th, 2011.

Old before their time.



TAKE a paper clip and straighten it out. Using just thumbs and fingers bend it in the middle to form a right angle. Then, at the same place, bend it back to form a right angle in the opposite direction. Do that half a dozen times or so and the paper clip will snap in two. The extraordinary thing about “metal fatigue” is that it takes only a few pounds of force applied repeatedly back and forth across the paper clip’s thickness to break it. To snap a typical paper clip in tension — by clamping one end and tugging on the other— would require a force of 50lbs or so.

The first to appreciate the catastrophic effects of stress-reversals were railway engineers in the 1840s. Broken axles caused countless accidents as railway lines crept across Europe and America. Being simply a rotating horizontal shaft with a heavy vertical load on it, early locomotive axles suffered severe stress reversals in their outer skins with every rotation. William Rankine, a Scottish engineer and one of the fathers of thermodynamics, was the first to explain how these repeated stress reversals could cause cracks to propagate. By the 1850s, the steam-engine pioneer James Braithwaite had coined the term “metal fatigue”.

The irony is that the lesson had to be relearned a century later. This time, aircraft manufacturers suffered the consequences. Their troubles began in the 1950s when they started flying higher and needed to pressurize the cabins of their passenger planes. Two de Havilland Comet aircraft—the world’s first commercial jet—broke up mysteriously in mid-air in 1954. Though it all but destroyed de Havilland, the disaster gave the industry crucial insights into how metal fatigue can rip an aircraft suddenly apart. It also taught them how to prevent stresses concentrating at certain points, thereby triggering a fatal tear in the aircraft’s skin.

Over the past few weeks, aircraft engineers have found they do not know quite as much about metal fatigue as they thought. The source of the problem that forced the Boeing 737-300 used on the Southwest Airlines flight 812 from Phoenix to Sacramento to make an emergency landing on April 1st, following a five-foot rent appearing in the upper-fuselage skin, has flummoxed engineers and safety officials alike.

By all accounts, it should not have happened. Admittedly, every time an airliner takes off and lands it goes through a demanding cycle of stress reversals. Like the paper-clip, the airframe and its alloy skin are stressed first in one direction as the cabin is pressurized while climbing to its cruising altitude, and then in the opposite direction when depressurized during descent for landing. On average, the short-haul aircraft used by Southwest, a budget carrier based in Dallas, do that half a dozen times a day — year in, year out.

The Boeing 737-300 in question was only 15 years old when its skin peeled open along a riveted lap joint while flying above 34,000 feet (just over 10,000 meters) with 118 passengers on board. The failure caused the cabin to lose pressure instantly and the oxygen masks to deploy. Within minutes, the pilot had got the plane down to 11,000 feet, where the passengers could begin to breath normally again. Shortly thereafter, the plane landed at a military base without further mishap or serious injury.

Much has been made of the 737-300’s age. However, a commercial aircraft that is 15 years old is still in its prime of life. The real issue is the way Southwest works its fleet so aggressively, specializing in rapid turnarounds. As a result, the plane concerned had accumulated nearly 40,000 flight cycles. An aircraft of that type and age would normally be expected to have logged little more than 30,000 flights.

Over the years, Boeing has probably accumulated more data on the fatigue life of airframes than any other plane-maker. After modifying the lap joints in the roof following some early failures, the company felt confident that its older 737s would be good for 60,000 cycles before they needed to be thoroughly tested for hairline cracks that could lead to fatigue failures. As an emergency precaution, the Federal Aviation Administration has now said that carriers operating 737s with the same lap-joint design along the roof should inspect the planes after no more than 30,000 cycles. Planes that have already logged 35,000 flights or more have to be inspected immediately.

What makes aircraft fatigue such a dark art is that, unlike standard tests done in a laboratory, an aircraft's structure has to endure a complex, mostly random, set of static as well as cyclical stresses when in service. Impurities in its material affect the fatigue life. So does the material’s hardness, and especially its surface condition. How the components were heat-treated in the factory is another factor. The operating temperature makes a difference, too. Worse still is the structural component’s shape: notches and sharp corners create concentrations of stress that can initiate cracks. The square windows on the original Comet jetliner were found to be the primary cause of its disintegration. Airliners have had windows with rounded corners ever since.

All things being equal, something they rarely are, the higher the cyclical stress level on an aircraft structure, the fewer the number of reversals it can withstand before breaking. As the stress level is gradually reduced, there comes a point where a structure can survive enough stress reversals to exceed the components expected life. By convention, the stress level that allows a component to survive 10m reversals is called its “endurance limit”. Unfortunately, the endurance limit is not some absolute— or even repeatable— value. When tested, identical samples can give widely different results.

In an ideal world, planes would be made of steel. Regrettably, that metal is too heavy for the job. However, steel alloys subjected to cyclic stress levels below their endurance limit rarely fail because of fatigue. In other words, they can be made to have an infinite life. The aluminum alloys that are used for their strength and lightness in aircraft construction are just the opposite. None can live indefinitely, and all will eventually fail from fatigue. Why that is so lies buried in the different crystalline structures of the two materials. That is something aircraft designers have to live with.

So what is an aircraft engineer to do? First, perform thousands of fatigue tests in the laboratory and then take a probabilistic view of things. Second, adjust the statistical results downward to account for differences between test conditions and the real world. Third, take in consideration all the known statistical variations of the material itself. The aim, as always, is to ensure that unpredictable factors do not reduce the fatigue life of a structure to less than that required.

Maddeningly, as Boeing has found of late, that is easier said than done. The biggest problem that aircraft engineers face is that the fatigue cycles a component faces are cumulative. No amount of resting can reset the clock. While an engineering student, your correspondent spent one whole summer crack-testing engine parts in an airline’s workshops in Spain. Parts that passed the stringent inspection were certified accordingly—and stored ready for reuse. What neither he, nor anyone else, could possibly know was whether the part in question would fail in the next 1,000 reversals or last for 10,000 more. No matter what Boeing does to rectify its aging 737s, their airframes are clocking up stress reversals that can never be expunged.

End of the article.

THE BOTTOM LINE IS: AVOID, BY ALL MEANS, TRAVELLING IN OLD BOEINGS 737, namely, models 737-300.

Dixi