More shuttles will be lost before 2020, study predicts

WASHINGTON — WASHINGTON - Even if NASA corrects the problem that doomed the Columbia, the agency is likely to lose more shuttles before the fleet reaches its planned retirement date of 2020, according to the draft of a study done for the Columbia Accident Investigation Board.

"Given uncertainties about the full set of causes for the loss of Columbia and given unforeseen and unplanned aging issues, it is not obvious that the shuttle system will reach even a 30-year life of useful service," says the draft, written by the Rand Corp. NASA hopes to fly the shuttles for 40 years.


This examination focuses closely on potential safety problems related to the shuttle fleet's age and is implicitly critical of NASA for not doing everything possible to monitor such problems as rust and metal fatigue.

The study, by Jean Gebman, an expert in aircraft aging, cites corrosion and fatigue of aging metal and parts, as well as the general level of reliability of the shuttles. It is based on an older Rand estimate of shuttle risks, updated with the experience of the Feb. 1 Columbia breakup.


The study, dated June 29, suggests that NASA needs to get a better grip on some potential age-related problems. For example, it says the agency never performed fatigue testing on the shuttles' wings although the wings are susceptible to cracking because they flex during liftoff, re-entry and transport.

Metal fatigue occurs with repeated flexing; the civilian aviation industry has been attuned to the problem since the roof of an Aloha Airlines Boeing 737 tore off during a flight in 1988.

Heating on re-entry can also cause strain on materials, according to the report. Even when the shuttle's thermal protection system works as designed, the study says, "different parts expand differently," creating internal tensions.

Gebman's draft quotes what appears to be a finding by the accident board - that NASA has not tested the wings or subassemblies for fatigue.

At NASA, however, Michael A. Greenfield, the associate deputy administrator for technical programs, said that when the agency was developing its shuttles, in 1979 and 1980, it used the Challenger to test the effects of flexing on the wings.

A wing of that shuttle was fitted with 2,000 sensors and bent back and forth to 110 percent of the load it was designed to take, he said. The design life of the orbiter was 100 flights, so the wing was tested to 400 flights, he said. (The accident that destroyed the Challenger on liftoff in 1986 was caused by a faulty seal in a rocket booster.)

"We know the fatigue properties of the materials - they are very well characterized," said Greenfield, who is co-chairman of NASA's "return to flight" program.

He said the thermal stresses had also been measured and matched the predictions of engineers during the design.


But Gebman, when he testified before the Columbia accident board in April, made a point of saying that fatigue in composite structures was not as well understood as fatigue in metals. This made an impression on Sheila Widnall, a member of the board. "I was afraid of that," she said at the time.

Gebman's draft also maintains that NASA has deferred inspections for corrosion, even though standing water had occasionally been found inside Atlantis (which is still available to fly) and Columbia after rainstorms. Columbia and Discovery each had corrosion behind the crew cabin, a location that is hard to inspect and to repair.

At one time, NASA had a "corrosion control board," but the study said it apparently no longer exists.

Al Feinberg, a spokesman for NASA, acknowledged that corrosion inspections and repairs had been deferred at times but said the corrosion control program itself still existed. Any postponement of an inspection or repair has been made "only after it has been assessed by engineers and been determined that it's minor and absolutely no safety issue," Feinberg said.

Neither the Challenger nor the Columbia accident appears to have been caused by aging parts. Columbia apparently broke apart on re-entry because of wing damage caused on liftoff by a chunk of foam thrown off the external tank.

The part of the wing that the foam hit, a leading edge panel, is known to be susceptible to degradation over time, and the accident board has already complained that NASA is not properly tracking the condition of those panels. But the impact was so severe that it would have damaged even a new panel, board members said.