posted 05-30-2003 11:21 AM
Mockup Wing Is Torn by Foam in Shuttle Test
New York Times 05/30/03
author: John Schwartz
author: Matthew L. Wald
c. 2003 New York Times Company
HOUSTON, May 29 — A piece of insulating foam shot at a mocked-up shuttle wing opened a long slit in its leading edge, which may help to explain what caused the loss of the space shuttle Columbia, investigators said today.
In the experiment, which was conducted for the independent board investigating the shuttle disaster, researchers shot a 1.67-pound chunk of foam from a gas cannon at a full-size model of the wing's leading edge at about 530 miles per hour. They were trying to recreate the circumstances at the Columbia's launching, when a piece of insulating foam from the external tank slammed into the shuttle wing at similar speed.
The impact produced a 22-inch-long gap, ranging in width from the thickness of a dime to a quarter inch, a spokesman for the Columbia Accident Investigation Board said this evening.
The test was not as realistic as possible; the target was made of fiberglass, not the material used in shuttles, reinforced carbon-carbon. Carbon-carbon, or R.C.C., is in short supply but will be used in a later test, said the spokesman, Lt. Col. Tyrone Woodyard of the Air Force. Fiberglass can withstand more forceful strikes without shattering, materials experts say. But investigators were excited by the test.
"If I was a gambling man, I'd bet it'll severely damage or perhaps even shatter a more brittle material such as the R.C.C.," said one of the commission's investigators. Another investigator said, however, that the main value would be to calibrate the testing mechanism, getting the speed right and the angle of impact, in this case 20 degrees, without using up scarce carbon-carbon samples.
In some ways, a narrow gap would be a more promising result than a shattered panel. Investigators note that before breaking up on Feb. 1, the Columbia entered the atmosphere far west of the California coast, but held together until it was over Central Texas, a sign that the breach in the wing must have been small and that the damage progressed relatively slowly.
That is why an expert outside the investigation suggested that today's experiment had solved the mystery. "That's the answer," said Paul A. Czysz, a professor emeritus at Parks College of Engineering and Aviation at St. Louis University, when told of the test results. A slit the size of one created in the test would let in a stream of gas three times as hot as a blowtorch."My God, that's like a barn door at those temperatures," he said.
Investigators have already concluded that a hole in the shuttle's left wing let in the superheated gases that destroyed the wing, and they knew that a piece of foam struck the wing on launching. But they would not have been able to link the two convincingly without experimental evidence, and some of them had been worried that the experiments might not produce any wing damage.
Adm. Harold W. Gehman Jr., who heads the investigation board, has repeatedly tried to lower expectations about the experimental results. He has said the board's recommendations will not be be based on absolute proof that the foam caused the hole.
Colonel Woodyard said the tests, which were conducted by the Southwest Research Institute in San Antonio, created a gap between a leading-edge panel and a part that fits between panels, called a T-seal for its shape. The impact appears to have shoved the seal sideways, exposing the inner structure of the wing behind it.
Professor Czysz suggested that the leading edge panels and seals shifted from the impact because the wing would not have been designed to resist a blow that exerted sideways pressure. "That just goes to show you that the thing you least expect to happen, will," he said.
A NASA engineer working with the investigators said the results were impressive but could not be conclusive until the tests were performed again using reinforced carbon-carbon. "All of the analysis and investigating and theorizing in the world just goes right down the tubes as soon as you have experimental information," he said. "The hardware doesn't lie."