THE computer model used to assess tile damage on the space shuttle Columbia was entirely inadequate, it emerged last week. 鈥淚t鈥檚 not really a computational model at all,鈥� says Hal Gehmann, the retired admiral who is leading the accident investigation. 鈥淚t鈥檚 essentially an Excel spreadsheet.鈥�
NASA used the software to determine that the shuttle could not have been seriously damaged when it was hit by a suitcase-sized chunk of foam that fell from the external fuel tank 82 seconds after lift-0ff. Wing damage from this impact is now thought to have led to the loss of the orbiter on 1 February.
As the Columbia Accident Investigation Board (CAIB) continues its investigation, the use of unsophisticated analysis software was not the only aspect of NASA鈥檚 post-launch behaviour that prevented a proper damage analysis.
Advertisement
CAIB member Sally Ride, a former astronaut says that the impact assessment began with a detailed analysis of a launch video to determine the size of the foam debris. Another group of engineers then used this information to work out where the piece might have hit the shuttle鈥檚 left wing. 鈥淭here was a large uncertainty in where it hit,鈥� says Ride.
This imprecise information was then passed on to NASA鈥檚 impact damage assessment group, who entered it into their computer model. This is essentially a spreadsheet-based look-up table that shows how much damage is produced by an impact with a piece of debris of a certain size. Although the shuttle鈥檚 delicate heat-shielding tiles were never designed to withstand impacts, NASA has built up a huge amount of data on it since almost every flight has suffered from some kind of impact damage 鈥� mainly from tank foam. It also fired pieces of debris into tiles to determine the damage this causes. The results were used to build the spreadsheet.
There was just one problem. 鈥淭he largest object they had fired at a tile was about 1 by 1 by 3 inches long,鈥� says Gehmann. Nowhere near as large as the object that struck Columbia. 鈥淪o when you go to look up where you are on the model, you鈥檙e not even on the same page,鈥� he says.
Ride, who interviewed the impact engineers, says they knew the limitations of their model and during a meeting to discuss the impact a few days after launch, asked NASA to obtain images of the shuttle in orbit so that a better assessment could be made. 鈥淭hey had good engineering reasons for wanting the imagery,鈥� she says. But the photographs were never taken 鈥� for reasons that are not yet clear. Ride wants to know why.
Meanwhile, the ground search for shuttle debris is drawing to a close. Recovery teams have found over 50,000 pieces of debris, roughly 32 per cent of the shuttle鈥檚 dry weight, which will be used in the reconstruction.
The CAIB is working on two interim recommendations as a result of its investigations so far. It is expected to recommend that NASA takes detailed photographs of all future shuttle flights in orbit 鈥� using US government satellites 鈥� and that it begins non-destructive testing of existing shuttle components to search for problems such as stress fractures caused by ageing.
The board is expected to begin writing its final report in June.