With scattered clues, teams re-create an unfolding calamity

Fusion of science, detective work pinpoints cause

By Gareth Cook, Globe Staff, 8/27/2003

As the shuttle Columbia glided homeward toward California, a jet of hot gas began to eat through the shuttle's left wing like a glowing worm, entering through a hole created by a piece of loose foam just after liftoff. As the Columbia flew on, its crew and controllers unaware, the 5,000-degree gas burst through the wing entirely and sent back a spray of molten metal. The wing started to sag back like a mattress on the roof of a speeding car. Finally, in the skies just southwest of Dallas, the wing gave way and the Columbia disintegrated, killing all seven astronauts.

In its 248-page report issued yesterday, the Columbia Accident Investigation Board left many lingering questions for NASA, such as how the space agency would handle the sweeping call for changes in its safety culture. But one point of certainty, say engineers, is the board's conclusion on the accident's physical causes.

"We should be satisfied that the technical issue has been resolved," said Wesley Harris, a professor at MIT and a former associate administrator for aeronautics at NASA.

In reconstructing the shuttle's flight, investigators faced enormous obstacles: The remains of the craft were either vaporized or scattered across the Texas backcountry, and a complacent NASA had done less than it had in the past to record the flight.

But with help from scientists in Boston and across the country, the board pieced together the details of the accident from an array of far-flung clues. The difficulty of the reconstruction was a key reason the investigation stretched to nearly seven months instead of the initially planned two, but board members saw it as essential.

"Think about TWA 800 -- the Martians did it, or the Navy shot it down," Sheila Widnall, a board member, said about the 1996 crash off Long Island and the variety of theories on its cause. Widnall is a former secretary of the Air Force and a professor of aeronautics and astronautics at MIT. "I believe the credibility of our report and the credibility of the recommendations we are making to NASA depend on us getting it 100 percent right."

The investigators considered five independent sets of evidence: the debris, imaging, the aerodynamics of the shuttle, the way heat travels around and through the shuttle, and the readings from on-board sensors -- particularly the way the sensors failed as hot gas severed wires on its way through the wing. All five sets of evidence pointed at precisely the same spot: a breach in the lower part of the left wing's leading edge, in "panel 8" of the tough reinforced carbon-carbon it is made of.

The investigation itself was enormous. The board had more than 120 staff members, the report says, and 400 NASA engineers were directly involved. More than 25,000 searchers walked vast stretches of wild land to recover a stunning 38 percent of the orbiter. Two of the searchers who worked for the US Forest Service were killed in a helicopter crash. The investigation also called on scientists around the country to solve difficult technical problems or track down loose ends. At the MIT's Lincoln Laboratory in Lexington, for example, a team of scientists was asked to identify a strange object that had apparently fallen off the shuttle. There was little information about the object -- just a few glimpses from radar.

For two months, about a dozen people at the laboratory worked intensively to identify the object. It was tedious going, a bit like trying to track a person's movements through Boston by scanning the backgrounds of all the photos taken by tourists over a few days.

Using snapshots of data from radars around the world, including the massive PAVE PAWS military radar array on Cape Cod, the team was eventually able to show the object was probably a curved piece from the leading edge of the shuttle's wing.

"It is not quite like being manic depressive, but it's close," said Grant Stokes, associate head of the aerospace division at Lincoln Lab. "We would lay all the clues on the table and then figure it wasn't quite right and break apart the puzzle and start again."

As the work progressed, a consistent picture began to emerge. The debris team reassembled the left wing in a Florida hangar, and there was an obvious hole around the panel 8 area of the leading edge. Near the hole were thick deposits of molten metal and insulating materials, unlike damage seen anywhere else, said Lieutenant Colonel Patrick Goodman.

Another team, headed by MIT's Widnall, used computers to model how the hot gases would have traveled through the wing and how the shuttle would have flown as the accident progressed.

A test done at the Southwest Research Institute in San Antonio, Texas, convinced any remaining skeptics. Engineers there fired a piece of foam at a section of shuttle wing, duplicating the precise characteristics of the liftoff mishap. The foam blew a large hole in the wing's surface.

The report issued yesterday included a list of 29 recommendations to make the shuttle safer, some of which dealt with the specific problems that led to the accident, such as preventing foam from breaking off and ensuring the wings can sustain foam strikes, as well as more long-term ones, such as an exhaustive reexamination of the entire shuttle design if it is to be flown past 2010.

Outside engineers said that it would be relatively easy to prevent this specific type of incident from happening but that it will be very difficult to keep NASA from drifting once again into complacency after the sting of the disaster starts to fade.

"What comes out of this more than anything else is the similarity between Challenger and Columbia," said John Tylko, a lecturer at MIT who follows the shuttle program closely. "You had a serious flaw in the shuttle design that had come to be accepted."

The more profound question, said many engineers, is how much risk Americans are willing to tolerate in their manned space program. If the program continues, even with dramatic changes, it is virtually certain that more astronauts will die.

"To get into orbit you have to go 17,000 miles per hour," said James Walker, a scientist at the Southwest Research Institute who helped in the investigation. "You have to be carrying that much fuel and you need to be going that fast. There is no getting around it."

Gareth Cook can be reached at cook@globe.com.