Space shuttle Discovery grounded for weeks by fuel leak, foam crack

The space shuttle Discovery, poised on the launch pad for its final mission, is acting like the little engine that didn't want to.

A leak allowing dangerous levels of hydrogen gas to build up around the orbiter has prompted mission managers to shift Discovery's launch from Friday to no earlier than Nov. 30.

The leak appeared about 7 a.m. Eastern Daylight Time as controllers were pumping fuel into the shuttle's external fuel tanks.

Engineers quickly spotted the problem and traced it to a fitting on the outside of the tank that bleeds excess hydrogen into a line that sends it to a flare stack roughly 0.2 miles from the pad, where the hydrogen is burned.

The same problem cropped up on two missions last year. This time around, the leak was worse, occurring when fuel was coursing into the tank at its highest rate. The concentration of hydrogen gas at the offending fitting was so high it exceeded the upper limit the leak sensors could measure.

As if to add insult to injury, when workers returned to the pad after controllers drained the fuel tanks, they discovered a seven-inch crack in foam insulation on the outside of the tank facing the shuttle, near the orbiter's nose. It looked as though the piece of foam involved had shifted slightly, which would have given ice a chance to build up in the area prior to lift-off.

The Columbia accident in 2003, in which seven astronauts were killed as the orbiter broke up on re-entry, was traced to foam and ice striking the orbiter's left wing during its ascent and damaging the tiles and carbon-composite materials designed to protect the underside of the shuttle from the heat of re-entry.

Since then, NASA has spent much time and effort to reduce the amount of debris the external fuel tanks shed, an effort that has largely been successful.

In describing the crack during a briefing this afternoon, Michael Moses, who chairs the prelaunch mission management team, said, "That's not something we like to see.

"I don't know if that would have passed our criteria to be 'go' for launch, but it certainly would have been something that would have generated a whole lot of discussion," he said.

The fuel leak, on the other hand, didn't require discussion. It clearly violated launch requirements as well as safety requirements for workers at the pad making final preparations for Discovery's lift-off.

Controllers faced a leak in a similar fitting in March 2009, another Discovery mission. By repeatedly opening and closing a valve inside the fitting, controllers were able to eliminate the leak. This suggested that a small piece of ice or other debris may have temporarily prevented the valve from closing.

But in June, the problem cropped up again. Cycling the valve between open and closed didn't stop the leak. Engineers ultimately traced the problem to ill-fitting seals at the fixture. Once new seals were installed and the fixture had passed its leak checks at essentially room temperature, controllers had to fill the tank as if for launch to see if the repair still held up under liquid-hydrogen temperatures, more than 423 degrees below zero F.

It did, clearing Discovery to launch a month after its initial lift-off date.

This time around, planners looked at the possibility of making repairs in time for a launch on Monday. The thinking was that the cause of a leak this large should be glaringly obvious compared with the two previous leaks and perhaps more straightforward to repair, said Michael Leinbach, the shuttle's launch director.

But as managers weighed their options, it became clear that a Monday launch wouldn't allow astronauts to remain on orbit the full 11 days planners had envisioned. Faced with the technical twofer they'd been handed – the hydrogen leaks and a foam problem – and the likelihood of an underachieving mission, managers opted to push the launch into late November at the earliest.

The postponement comes on the heels of a week of launch delays – first to fix leaks in systems that provide gas to pressurize fuel tanks that feed the shuttle's on-orbit maneuvering system, then to fix problems with a circuit breaker that routes electricity to systems that control the shuttle's main engines. A third delay came Thursday because of bad weather at the pad.

Looking back at the week's glitches, Mr. Leinbach acknowledged that "it gets frustrating because we'd rather be launching. But it's a machine, and every now and then machines break. Our machine is broken and we need to go fix it."