It looked like a flawless landing. Two days after Christmas 1999, a little after 7pm, the space shuttle Discovery touched down safely at the Kennedy Space Centre in Florida, its eight-day mission to service the Hubble space telescope complete.
It was another fortnight before a Nasa engineer spotted something unexpected. Running across a reinforced panel on the leading edge of Discovery’s left wing was a fissure, a crack so deep that the ‘substrate’, the weaker material behind the panel, was visible.
Unsure what had caused this potentially lethal problem, Nasa did nothing. The official ‘inflight anomaly report’ filed by the engineer records no resolution. A couple of flights later, when Discovery went back into orbit to visit the International Space Station in March 2001, the same thing happened again.
On this occasion, Nasa ordered its engineers to check the wings of all four shuttles. They were not given any specialised equipment, but told to look for cracks with their naked eyes and to prod the vital leading edges with their fingers, to look for soft spots. Another crack was in the wing of the shuttle Atlantis.
On 1 February 2003, investigators believe another space shuttle had a cracked left wing. Perhaps the fissure was a little wider; this time, Nasa’s luck ran out. As Columbia, the oldest vessel in the shuttle fleet, began to descend into the atmosphere at a speed of five miles per second, a jet of ionised air burnt through the crack, through the substrate, into the wing, incinerating control and sensor wires and melting metal. Debris from the craft fell across the southern United States. The seven astronauts aboard had no chance.
Now, as the Columbia disaster Accident Investigation Board under Admiral Harold Gehman prepares its report for publication at the end of next month, The Observer has learnt the previous cases of wing damage, and the underlying reasons for Nasa’s failure to heed their ominous warnings, have emerged as two central issues.
Former top Nasa engineering and safety staff in close touch with Gehman’s team say the accident was preventable. Deep cuts in the shuttle’s budget when the Bush Administration took office deprived it of essential upgrades, leaving safety and inspection procedures outmoded and inadequate.
‘We are talking about a systemic failure,’ said Professor Henry McDonald, former director of Nasa California and the chairman of a wide-ranging space shuttle inquiry team which reported in March 2000. ‘The mentality became, ”We flew it, we had a problem, we landed, so what’s the big deal? So there was a crack in the wing panel, but hey, we got away with it; when’s the next mission?”’
With deep budget cuts, said Richard Blomberg, former chair of Nasa’s watchdog, the Aerospace Safety Advisory Panel, pressure on Nasa became unbearable: ‘to do more with less, to make sub-optimal decisions one could not be happy with’.
McDonald and Blomberg were forced to leave Nasa less than a year before Columbia’s final flight: Blomberg in April 2002, McDonald last November. Other senior figures involved in the shuttles have also gone, including Nasa’s boss, Dan Goldin, and his Associate Administrator, Joe Rothenberg.
‘I cannot speak for the others,’ Blomberg said. ‘But one could speculate I was removed because I was saying things that some people found uncomfortable. It was becoming increasingly hard to say what had to be done on safety grounds because the programme was on shaky grounds financially.’
It was not meant to be like this. After the disaster in 1986, when Challenger exploded soon after launch, the official inquiry found that knowledge of the cause, a tendency of fuel tank seals to become brittle and leak in cold weather, was widespread. Sweeping reforms were promised.
But an effective budget freeze for most of the 1990s saw those aspirations decay. Nasa’s workforce was halved, and many of its operations privatised. In July 1999, Columbia came close to exploding when fuel leaked and two of its engine controllers ceased to function after a short circuit. In response, McDonald’s independent assessment team was asked to report on all aspects of the shuttle and its management.
When Columbia was stripped down, his team found almost 4 000 potentially lethal anomalies in 240 kilometres of wiring insulation, most of it made of a material called Kapton which gets brittle with age, and has been blamed for several civil aviation disasters. McDonald also highlighted other shuttle components as desperately in need of upgrades.
His analysis of the shortcomings of the shuttle programme’s organisation was even more significant. Noting grimly that the shuttle flew in a ‘one-strike-and-you’re-out environment’, it found defects in Nasa’s test equipment and methods, and noted that its system for identifying problems was ‘rudimentary’.
Databases could not be searched for recurrent flaws, and there was no way of alerting staff when they cropped up. There was no automated system for flagging up the fact that the spacecraft had a tendency to develop cracks in the leading edge of its wings. In the Columbia crash, all these themes were to become tragically apparent.
Gehman’s report seems certain to find that the fatal crack was made by a 1,5kg lump of foam, part of the external fuel tanks’ protective covering, which hit the wing at 500mph during the climb to orbit. Nasa has been aware of this problem for years but because it has happened numerous times — McDonald said up to 100 such ‘dings’ from foam and other tank debris were considered normal — it was not thought dangerous. Nasa considered 100 dings per mission safe.
The 100-ding average was based on counting dents in the heat-resistant tiles covering most of the shuttle’s body. Relatively soft, their position and alignment mean a pool of ‘dead air’ collects in a dent during re-entry, so the shuttle’s heat shield remains little compromised.
What Nasa forgot, and what the fault-tracking system failed to tell it, was if staff were counting dents in the tiles, it was certain the wings’ leading edges were getting bumped as well. Made of panels of reinforced carbon-carbon, or RCC, these edges are the shuttles’ most vulnerable point in the white heat of re-entry.
Hard as granite, they don’t show dents, but blows may weaken it catastrophically: ‘The fact that 100 dings were considered acceptable was in no way an adequate assessment of the real level of risk,’ said McDonald.
The only way to tell if RCC is safe is through what engineers call non-destructive evaluation, NDE – techniques such as ultrasound which allow them to see beneath the material’s surface.
A programme of NDE was one of the 2000 McDonald report’s key recommendations. Like many of the others, having first been accepted by Nasa, when the Bush Administration cut the shuttle upgrade budget from $1.6 billion over five years by almost half, it fell by the wayside.
Sean O’Keefe, appointed Nasa Administrator last year, was largely responsible for slashing the agency budget in his former incarnation as deputy chief of the Administration’s Office of Management and Budget. In a recent interview, he claimed there was ‘no effective technology right now’ to perform NDE. Nasa spokesman James Hartsfield added that visual inspection and finger-prodding was a good method of detecting flaws.
McDonald disagrees. ‘It’s obviously not an effective test. If you found a soft patch, the panel would already have gone way beyond the point of safe operation.’
He is contemptuous of O’Keefe’s suggestion that one of the earlier Discovery cracks had been caused by rainwater corroding paint. ‘As an engineer, I would want much more explanation. RCC is not a material subject to corrosion.’
Without NDE, he said, the only safe way to treat RCC panels is to throw them away after a finite number of missions.
For now that is not going to happen. Former safety panel chair Richard Blomberg reveals that not only did Nasa have no method of NDE, it had almost no spare RCC panels, and nowhere to make new ones – which is why lab tests for the Gehman investigation have had to be done on panels cannibalised from Discovery.
‘Shown details of the previous wing crack incidents, Blomberg proves McDonald’s point about fault reporting: ‘It should have been brought to the panel’s attention. It did not come up to us. I’m surprised we weren’t told, but this is the first I’ve heard about this.’
That failure appears to reflect a deeper malaise: ‘The pressure was always to get the next flight off. There was a faction in Nasa that thought the panel should be less independent, more of a consultative body instead of a wide-ranging oversight group.’ That faction, he said, became stronger with the 2001 change of administration, and this was ultimately the reason he was fired.
Nasa officials failed to return The Observer’s calls but its website has a new section: details of the shuttle ‘return to flight task force’ — plans to get the dwindling fleet back in orbit by the end of the year. – Guardian Unlimited Â