The New York Times February 28, 2003
NASA Pressed on When Officials Learned of E-Mail About Shuttle
By Kenneth Chang and Richard A. Oppel Jr.
NASA's top official faced sharp questioning yesterday over when mission managers learned of e-mail discussing dire possibilities facing the space shuttle Columbia.
Lawmakers questioned Sean O'Keefe, the NASA administrator, about why he learned only on Wednesday of e-mail messages in which engineers discuss whether chunks of foam insulation that hit the shuttle during its ascent might cause catastrophic failures in the landing gear and hydraulics. Representative Anthony Weiner, a New York Democrat, asked why Mr. O'Keefe didn't learn earlier of "this vigorous debate among experts" at the space agency. "Have you fired anyone for not bringing them to your attention sooner?"
Mr. O'Keefe said concerns about the shuttle's condition were properly examined by NASA officials during the mission, but he said he would await the judgment of an independent investigative board "as to whether that was an appropriate systemic or management approach."
The hearing, held in Washington before the House Science Committee, broadens what had been a technical investigation into a wider questioning of NASA's management. As they were after the Challenger explosion in 1986, NASA officials are again being asked whether failures in management prevented safety concerns from being addressed.
"It's not only the data and what has happened to Columbia but how NASA is operating," said Dr. M. Elisabeth Pate-Cornell, chairwoman of management science and engineering at Stanford University and an author of a 1990 report that concluded foam debris could seriously damage the shuttle's heat-resistant tiles.
"They also have to modify their culture to accommodate uncertainty and the notion it's not because you have survived something in the last few months that you're necessarily going to survive it in the next flight," Dr. Pate-Cornell said. "They have to do their analysis with a lot more consideration of the uncertainties to better support the decision-making."
Just last week, Mr. O'Keefe said there was a "scant bit of information" that indicated anything troubling was amiss until eight minutes before mission control lost communications with the shuttle. "There was absolutely nothing that registered on the scale," he said.
But the stream of NASA documents released in recent days shows that rather than being inexplicable and unforeseen, the destruction of the space shuttle Columbia appears to have unfolded as a logical chain of events, almost exactly as NASA's experts and diagnostic tools had predicted during the flight.
One of those tools, a computer program, calculated that the falling foam from the external fuel tank during launching would have destroyed a swath of tiles. Given the assumption of extensive tile damage, NASA engineers deduced that on re-entry the wing would have burned through, leading to a cascade of problems almost identical to those Columbia experienced on re-entry.
NASA engineers did not act on the warnings because engineers at Boeing, a major space shuttle contractor, assumed that a quarter-inch of tile would survive regardless of the size of the impact and concluded that there was no serious danger to the shuttle.
If that analysis was overly optimistic, and the impact had exposed the underlying aluminum, "an entirely coherent explanation of what went wrong" emerges, said John Pike, a space industry analyst and director of GlobalSecurity.org. "The pieces pretty much fit together," he said.
NASA is now trying to explain these issues:
*Why it quashed a request to the United States Space Command for help inspecting Columbia's underside, even before it had estimates of the extent of damage.
*Why Boeing engineers discounted predictions of extensive destruction of the tiles by the computer program specifically designed to assess damage from debris impacts.
*How closely NASA engineers scrutinized the Boeing analysis.
*Whether the last-day speculation about potential malfunctions with the landing gear was conveyed to mission managers and what, if any, contingency plans the managers made in case the crew had to bail out or make an emergency landing.
Representative Sherwood Boehlert, the New York Republican who is chairman of the science committee, defended Mr. O'Keefe, describing the e-mail messages released on Wednesday as speculation from engineers, and he urged lawmakers "to avoid simplistic comparisons" to efforts by NASA officials to cover up problems in the aftermath of the Challenger shuttle explosion in 1986.
But he asked, "Isn't the decision to cancel an inspection the kind of decision the administrator might be expected to be involved in?"
Mr. O'Keefe replied that officials determined that "the level of imagery" from the photographs "would not have been high enough to make a determination that everybody thought would be necessary."
"Was that a judgment call that was in error? We'll find out," he said. Asked whether the inspection decision should have been made at the "highest levels of NASA," Mr. O'Keefe replied, "Could be."
After the hearing, Mr. O'Keefe said he did not know who at NASA canceled the inspections.
On Wednesday, at Johnson Space Center in Houston, two authors of the series of e-mail messages spoke briefly to reporters. "This was just a mental exercise that we went through to 'what if' the whole thing," said Jeffrey V. Kling, a flight controller. On Jan. 31, he sent an e-mail message that said if hot gasses leaked into the wing, damaging the landing gear, he advised that the crew bail out "assuming the wing doesn't burn off before we can get the crew out."
Mr. Kling said that the ground controllers did not relay their thinking to the shuttle crew "because we did not expect it to happen."
Investigators have not yet ruled out other explanations for the disaster, like a strike by a micrometeorite, radiation from a solar storm, but no evidence to support them has surfaced.
However, the stream of e-mail messages, released in recent days under requests made by news organizations, show that engineers at NASA had considered in some detail the possible consequences of damage to tile from the foam but concluded they were highly unlikely.
On Jan. 17, a day after liftoff, a review of videos of the launching revealed that a piece of debris, apparently insulating foam that fell off the external fuel tank, slammed into Columbia's left wing 82 seconds into flight.
Examining the quickly moving white dot on the videos, a team of engineers at Boeing estimated that the size of the debris -- up 20 inches by 16 inches by 6 inches -- and its speed at impact -- 440 to 500 miles per hour. If the debris were foam, it would weigh about 2.67 pounds. They also calculated that the debris hit the underside of the left wing, on or near the wheel door. They presented their findings on Friday, Jan. 21.
A second Boeing team then calculated how deep a gouge a falling piece of foam would rip through the fragile ceramic tiles. Using a computer program called Crater written for this task, they considered nine possible trajectories for the debris. Crater calculated gouges from 19 inches long and 2.4 inches wide to 32 inches long and 7 inches wide. In each case, it predicted that the gouge would dip deeper than the thickness of the tiles on the left wing.
The engineers, however, assumed that a "densified layer" at the base of each tile, about a quarter of an inch thick, would survive the impact and still provide some protection against the heat of re-entry, James Hartsfield, a spokesman at NASA's Johnson Space Center, said in an interview yesterday. He said he did know if the engineers had done any analysis of whether the impact could have also scraped off the densified layer, exposing the shuttle's underlying aluminum wing.
At a news conference on Feb. 5, Ron D. Dittemore, the space shuttle program manager, discounted the foam as the cause of the accident. He said that engineers had, in the computer simulations, doubled the speed of the foam to more than 1,000 miles per hour - which represents a quadrupling of the kinetic energy. Yet the results still indicated no serious issues, Mr. Dittemore said. "And so it's difficult for us to believe as engineers, as management, and as a team, that this particular piece of foam debris shedding from the tank represented a safety of flight issue," he said at the news conference.
Mr. Hartsfield said yesterday that the Boeing engineers did not double the velocities in their analysis and that Mr. Dittemore had misspoken.
"My big curiosity is why NASA has so much faith in that Boeing report," said Paul S. Fischbeck, a professor of social and decision sciences at Carnegie-Mellon University who wrote the 1990 NASA report with Dr. Pate-Cornell. "That's the one study that seems to have been done that doesn't match what happened."
The Boeing engineers made their presentation to mission management team on Jan. 24. The slides from that meeting do not include the Crater results.
"They think their analysis was sound," said Edmund Memi, a Boeing spokesman. "They still stand by that."
Dr. Fischbeck said with the Boeing study was a "bottleneck" to more serious "what if" discussions of what they should do if problems arose during re-entry and landing. "They had a model that predicted how much damage would be done, but they discounted it, so they didn't look beyond it," he said. "They didn't seriously consider any of the outcomes beyond minor tile damage."
GRAPHIC: Photos: A piece of underside of the Columbia's left wing is among the debris being studied by investigators.; This image of the space shuttle Columbia's left wing was captured during the craft's fifth orbit. (Photographs by NASA)(pg. A20)
Chart/Diagram: "EVALUATING THE SHUTTLE: Clues About Columbia's Demise"
Assessing the impact of the foam that struck Columbia's left wing, a computer program called Crater calculated that the foam would have destroyed a swath of tiles. Engineers analyzing the tile damage before re-entry, later declared that there were no "safety of flight" issues.
Looking at the possibility of extensive tile damage exposing the underlying aluminum, experts had discussed a chain of events that could lead to the destruction of the shuttle.
Impact Estimated by Crater
Diagram of left wing highlighting the locations of the following items:
PREDICTED
IMPACT AREA
IMPACT GOUGES
EXCEED TILE DEPTH
LEFT MAIN WHEEL WELL
LEFT WING LEADING EDGE
Sensor Readings on the Flight
Diagram of the left wing highlighting the location and sensor readings for the following items:
OFFLINE
HIGH TEMPERATURE
Brake line temperature
Fuselage temperature
Tire presssure
Wheel temperatures
Landing gear hydraulic temperatures
JEFFREY V. KLING, flight controller at the Johnson Space Center
"If there was hot plasma sneaking into the wheel wells, we would see increases in our landing gear temperatures and likely our tire pressures."
KEVIN McCLUNEY, flight controller at the Johnson Space Center
"Let's surmise just what sort of signature we'd see if a limited stream of hot plasma did get into the wheel well . . . . Tire pressure (and theoretically brake pressures as the fluid temperature increased, though the expansion is small) would rise given enough time . . . . Then the data would start dropping out as the wiring is severed."
(Sources: NASA; Boeing)(pg. A20)
Copyright © 2003, The New York Times Company
