Boeing Discounted Analysis of Shuttle Damage
Boeing Co. engineers may have failed to recognize that the Columbia had been crippled by a liftoff accident because they arbitrarily dismissed the results of a computer program designed to assess the integrity of the craft, NASA investigators said Tuesday.
The computer program, known as Crater, predicted that the space shuttle suffered “significant” damage when it was struck by a piece of foam insulation shortly after lifting off from Cape Canaveral, Fla., on Jan. 16.
The Boeing engineers, asked by NASA to assess the damage caused by the piece of insulation, appear to have overruled the program because they believed that it was too conservative and was overstating the potential damage.
Members and representatives of an independent panel probing the Columbia disaster said Tuesday that they plan to investigate whether the decision was based on subjective judgment rather than sound science.
Officials say it is too early to say if the Boeing analysis was wrong or whether the liftoff incident even had anything to do with the Columbia’s Feb. 1 disintegration. But the panel is worried enough that it might order another round of tests on the potential damage caused by insulation that fell from an external fuel tank and struck the shuttle during liftoff.
The tanks, which are filled with liquid oxygen and liquid hydrogen and are used to carry the space shuttle to orbit, fall back through the atmosphere after they are jettisoned. The tanks break into pieces on the way down. Many burn up in the atmosphere. The rest fall into the sea, and NASA makes no attempt to recover them.
The new tests, officials said, would likely be conducted on a fuel tank identical to the one that helped carry the Columbia into orbit. That tank, by coincidence, was recently built at a Lockheed Martin Space Systems Co. plant in New Orleans. The Columbia Accident Investigation Board has impounded the tank, the board’s chairman, retired Navy Adm. Harold W. Gehman Jr., said Tuesday. “Nobody’s touching it,” he said.
Because the investigative board might have to destroy the impounded tank in order to test it, “we only get one shot at it,” Gehman said. Officials said they have not determined how they will use the external tank.
Boeing declined to answer questions about its analysis Tuesday. Dan Beck, a Boeing spokesman, said that “NASA would rather we didn’t talk about it,” but he said the company continues to stand behind its work. One of the five engineers who wrote the report, reached Tuesday in Houston, said he had been instructed “not to answer any questions.”
The concerns center on an incident that occurred about 80 seconds after the Columbia lifted off. Material believed to be a 2.7-pound piece of foam insulation popped off an external tank and struck the left side of the shuttle -- the same portion of the craft where sensors detected a series of failures in the minutes before the disintegration. NASA engineers discovered the incident while reviewing film of the liftoff the day after launch.
Five days into the mission, NASA asked engineers at Boeing, a prime contractor for the space agency, to determine whether the incident was cause for alarm. The answer, in short, was no.
A copy of Boeing’s analysis indicates that its engineers relied heavily on Crater, a NASA-designed computer program. Crater is based on a database developed from years of tests conducted on spaceships, particularly on the heat-resistant tiles that cover much of the shuttle, protecting it from the intense temperatures encountered while reentering the atmosphere. In recent years, NASA engineers have fired everything from ice to BBs to .22-caliber bullets at those tiles, and have entered many of the results into the program.
In the Boeing analysis, Crater predicted “significant tile damage” and said the piece of foam insulation, which was traveling more than 500 mph, might have penetrated more than three inches into portions of the space shuttle protected by tile that was less than two inches thick. Some analysts believe that could have exposed sizable sections of the shuttle’s aluminum shell to the heat of reentry.
Aluminum melts at about 1,000 degrees; the superheated gas known as “plasma” that is thought to have penetrated the craft seconds before it broke apart might have been as hot as 3,000 degrees.
The Crater program does not appear to have specifically predicted that the space shuttle would be destroyed upon reentry.
In its report to NASA, the Boeing engineers wrote that the program was “designed to be conservative” and often “reports damage for test conditions that [in live testing] show no damage.” The computer program also “overpredicted penetration” of the protective tiles, the engineers wrote.
On Jan. 23, the Boeing engineers, apparently discounting dire scenarios envisioned by the computer program, predicted a “safe return.” Nine days later, plasma blasted into the left side of the space shuttle, possibly in the wheel well where its landing gear is stored during flight. The shuttle began breaking apart as it soared about 200,000 feet above California, then disintegrated over Texas, killing its seven-member crew.
A member of the investigative board said Tuesday that he would like to know what basis Boeing engineers had for discounting the results of the Crater program. Steve Nesbitt, a spokesman for the investigative board, confirmed that its members want to “hear more about” the methodology used in the Boeing analysis.
“[Boeing] concluded that Crater was conservative based on a very small number of samples. That was going a bit fast,” said M. Elisabeth Pate-Cornell, chairwoman of Stanford University’s Department of Management Science and Engineering and the principle investigator for a 1990 study assessing the risks of tile damage. Pate-Cornell has reviewed the Boeing analysis.
Engineers and NASA investigators said they fear the problems with the analysis only start there.
Pate-Cornell said in an interview Tuesday that she and other engineers turned up evidence 13 years ago that some of the bonding agents used to attach the tiles had failed during past missions. If Boeing was truly trying to assess the worst-case scenario for NASA, Pate-Cornell asked, why didn’t its engineers factor in the possibility that the foam insulation had not only struck the shuttle, but had struck tiles that were glued on poorly?
“That attracted my attention immediately,” Pate-Cornell said. “Where was the possibility of tiles that had not been well-bonded? They claimed they were using the worst-case scenario [to form their conclusions]. I don’t think that’s true.”
James Hallock, a member of the investigative board, also said in an interview Tuesday that he would like to know why Boeing does not appear to have weighed the possibility that the foam insulation that struck the shuttle might have been laden with ice. That would have made it heavier and increased the damage it might have caused.
What’s more, board members said Tuesday, investigators have now widened their probe of the liftoff incident to include the possibility that the object that struck the shuttle wasn’t a piece light foam from the tank’s outside layer but a heavier piece of internal coating material.
Steven Wallace, director of the Federal Aviation Administration’s Office of Accident Investigation, said that material is similar to felt and is about half an inch thick.
“Was it the foam?” Hallock asked. “Or was it the material behind the foam? Or was it ice?”
Hallock, the U.S. Department of Transportation aviation safety division chief in Cambridge, Mass., also said that the Boeing analysis appears to be founded on tests conducted with a 3-cubic-inch piece of foam insulation. NASA believes the insulation that struck the Columbia was 1,920 cubic inches.
“No one really tested anything that big,” Hallock said. “We’re looking at these types of things.”
Engineers say they were startled when NASA Administrator Sean O’Keefe, appearing before congressional investigators in Washington, compared the potential impact of the foam insulation to that of a foam cooler that falls off the back of a truck and hits a car going 50 mph.
“It’s not like that at all,” said Paul Czysz, professor emeritus of aerospace and mechanical engineering at St. Louis University and a veteran NASA consultant. “It’s like the shuttle got hit with a sledgehammer -- a bunch of sledgehammers.”
NASA has since said that O’Keefe misspoke.
*
Times staff writer Ralph Vartabedian from Los Angeles contributed to this report.
