Twenty-nine months ago, a loosened piece of fuel-tank foam insulation weighing 1.67 pounds struck the wing of the ascending space shuttle Columbia at a speed of more than 500 mph, precipitating the destruction of the vehicle and its crew when it reentered the atmosphere nearly two weeks later.
The disaster on Feb. 1, 2003, gave rise to an extensive effort by NASA to avoid another instance of "foam shedding" from the insulation that is sprayed on the fuel tank to keep its volatile hydrogen and oxygen fuel in a supercooled, liquid state. Technicians were retrained, and some foam was simply eliminated.
The loss of a substantial chunk of foam during Tuesday's launch of the shuttle Discovery starkly demonstrated that this effort failed. But it was hardly a surprise.
A task force established by NASA to monitor the agency's safety improvements after the 2003 disaster concluded on June 28 that the agency "did not meet" the requirement of a team of accident investigators that the agency "eliminate all . . . debris shedding." It noted pointedly that the external fuel tank, attached by metal struts to the shuttle during its violent launch, "still sheds debris that could potentially cripple an orbiter."
A crippling event apparently did not occur Tuesday, but the recurrence of substantial foam shedding caused NASA to recalculate its estimate of the risks of continuing to fly the shuttles without first trying harder to fix the foam problem.
Although the accident investigators -- formally known as the Columbia Accident Investigation Board -- blamed the 2003 disaster on institutional and scientific lapses at NASA, they noted the challenge of trying to understand the cause and the impact of foam shedding during the moments when the external tank and the shuttle vehicle are operating "near the limits of their performance."
Moreover, because about 10 percent of the foam must be applied manually -- the rest is done by machines -- it can never be precise or perfect.
Since the Columbia accident, the agency has conducted hundreds of tests with foam "bullets" to assess the vulnerability of the shuttle to debris. NASA stripped the inch-thick foam from struts from which it had shed repeatedly, and the agency conducted mock, filmed trials of new foam-spraying techniques to minimize gaps in which air could collect and expand during the shuttle launch.
NASA also prepared for possible problems by upgrading its camera systems to observe the launch sequence, putting into place plans to inspect the shuttle's belly for debris damage after launch, and equipping the crew with a rudimentary in-flight repair kit -- all because of the troublesome foam.
"Discovery's [fuel] tank has been redesigned over the course of the last 24 months through testing and implementation of improvements that eliminate the chance of shuttle-damaging foam coming off during launch," the agency said in a news release before the launch this week. "It is undoubtedly the safest, most reliable tank ever built."
NASA nonetheless experienced familiar lapses as it attempted to cure the problem in preparation for Discovery's flight, according to the Return to Flight Task Group, staffed by space engineers and by retired military and NASA officials. Specifically, it modified the tank to reduce the shedding but used the wrong data in calculating how big an improvement was needed.
The task force blamed the mistake on "a lack of rigor in both development and testing." The public "executive summary" of its final report did not provide further detail.
NASA may also have bet wrongly in the final months before the launch that external tank ice -- rather than foam insulation -- posed the largest risk to the shuttle. It delayed the launch twice primarily to confront the ice problem, but in the end, it was the foam that vexed the agency again.