杏吧原创

NASA should not rely on its current computer models to clear the next shuttle for launch, according to an oversight panel. However, the panel still expect NASA to return to flight in May or June 2005, as planned.

The space agency has focused its return-to-flight efforts on reducing the amount and size of foam insulation that can break away from a shuttle鈥檚 external fuel tank. A briefcase-sized chunk of foam doomed the shuttle Columbia when it punched a hole in the leading edge of its left wing just after launch in January 2003.

Since the disaster, the agency has been conducting tests to find out why the foam comes off the tank, how the foam flies through the air and the minimum energy it must have to damage the shuttle鈥檚 鈥渢hermal protection system鈥�, including the reinforced carbon panels covering the wings鈥� leading edges.

NASA has used that minimum energy to model the maximum speeds that particles of various sizes could reach during a flight. Its estimate for the mass of the smallest piece of breakaway foam that could seriously damage the shuttle ranges from 10.5 grams at the top of the external tank to 34.0 g nearer the bottom. The estimate is lighter at the top because foam there has farther to fall to the shuttle wing, meaning it has a greater impact velocity.

鈥淔aulty assumptions鈥�

But the current models are incomplete and 鈥渃annot be used to precisely determine 鈥榓llowable鈥� debris鈥�, writes the independent Return to Flight Task Group in an interim report released on Friday. The 26-member panel is charged with assessing whether NASA has implemented 15 recommendations made by the Columbia Accident Investigation Board (CAIB).

The group says the Columbia disaster was caused by NASA making the 鈥渇aulty assumption鈥� that foam could not damage the shuttle鈥檚 wings. Now, the group writes that it is 鈥渃oncerned that NASA does not replicate the reliance on faulty assumptions and the results of analytical models to justify return to flight鈥�.

It says NASA has not yet demonstrated the 鈥渞igour鈥� of its mathematical models and that, until it does, the models should not be used for the certification of flight hardware or risk assessment. Unvalidated models can give a 鈥渇alse sense of risk reduction鈥�, said ex-astronaut Richard Covey, the group鈥檚 co-chair.

Commercial aircraft

But John Muratore, manager of systems engineering and integration for the space shuttle program at NASA鈥檚 Johnson Space Center in Houston, Texas, US, says: 鈥淲e have validated the models as best we can.鈥� He says NASA based its estimates on four independent models of how foam travels and two of how it can damage the thermal protection system.

鈥淏ut in the end, we can鈥檛 go out and fly shuttles at Mach 2 to fully validate them,鈥� he told New 杏吧原创. He notes that commercial aircraft must fly hundreds of times to acquire enough flight test data to show they are safe. 鈥淲e鈥檝e used wind tunnel tests and ground tests, but we really haven鈥檛 been able to validate by experiment in the way you would a commercial aircraft.鈥�

Despite its reservations about the modelling, the task group believes it is 鈥渢he safest tank that will have flown鈥�, says Covey. He says NASA has completed about 90% of the work required for the task group to make its final assessment, planned for April.

Nonetheless, eight of the 15 CAIB recommendations remain 鈥渙pen鈥� 鈥� including the 鈥渓argest remaining concern involving the ability to detect and repair damage to the shuttle while in orbit鈥�, according to the report.

But while the task group has debated the intent of that recommendation, Covey does not think it is a 鈥渟howstopper鈥� that could prevent the shuttle Discovery taking off as planned in May or June. NASA will probably fly with some repair capability and continue to improve this for future flights.