Shuttle Discovery launch postponed once again
BY WILLIAM HARWOOD
STORY WRITTEN FOR CBS NEWS "SPACE PLACE" & USED WITH PERMISSION
Posted: February 13, 2009
Launch of the shuttle Discovery on a space station assembly mission, already delayed twice because of concern about suspect hydrogen valves, was pushed back again late today, to no earlier than Feb. 27, to give engineers more time to assess the safety of the shuttle's external tank pressurization system.
NASA managers made the decision at the end of a marathon meeting to discuss test results to date and how those results might fit into a rationale to launch Discovery as is. The alternative - a possible valve redesign or modification - likely would trigger a lengthy delay with potentially significant downstream impacts to space station assembly.
While testing to date has gone well, sources said engineers have not yet been able to reach any major conclusions or even decide on exactly what triggered a failure observed in the most recent shuttle flight. It now seems clear that tiny cracks in the valves in question are a common phenomenon, not rare as previously believed, but only one valve has ever cracked to the point that a piece actually broke away.
In any case, shuttle Program Manager John Shannon decided late today to delay a second flight readiness review from next Wednesday to next Friday, pushing launch to no earlier than Feb. 27.
"The team just needed more time to bring in all of the data, basically," said NASA spokesman Kyle Herring at the Johnson Space Center in Houston. "They've done quite a bit in terms of the progress they've made, significant progress, but there's still some open issues associated with filling in some of the test data. They've still got some data coming in even as the meeting was going on. Shannon just basically decided the best thing to do was give everybody the weekend to put their thoughts together and put the rest of the data together and that just wasn't going to fit a Wednesday flight readiness review."
A NASA statement said the new target date "is not expected to affect the launch dates for missions that will follow Discovery's flight, STS-125 to NASA's Hubble Space Telescope and STS-127 to the International Space Station."
Herring said the engineering community expressed confidence "they could get all the test data in, build the computer models and bring all of that to the flight readiness review next Friday."
What happens after that is anyone's guess.
Discovery originally was scheduled for launch this week - Feb. 12 - on a mission to deliver a final set of solar arrays to the international space station. But the flight was delayed, first to no earlier than Feb. 19 and then to Feb. 22 because of concerns about the integrity of three critical hydrogen flow control valves.
The valves work like pop-up sprinklers, extending as required during launch to provide hydrogen gas to the shuttle's external tank to keep the tank properly pressurized. If a tank is over pressurized, hydrogen would be dumped overboard through a relief valve, putting the shuttle in a potentially catastrophic environment. An under-pressurized tank could lead to premature engine shutdowns.
The valve issue cropped up after the shuttle Endeavour's flight last November. During launch, telemetry indicated one of the ship's three hydrogen flow control valves allowed more hydrogen gas to pass through than expected. The shuttle's flight computers compensated by having the other two valves deliver less to maintain the proper tank pressurization.
After Endeavour returned to Earth, the suspect valve was removed and inspected. Engineers quickly discovered that a small part of the valve's lip, or poppet, had broken away, providing a pathway for the additional hydrogen gas. This was the first such incident in 124 shuttle flights.
In the wake of the discovery, all the FCVs on Endeavour, Discovery and Atlantis were removed and inspected, using dye penetration tests to look for signs of cracks. The valves also were subjected to scanning electron microscope inspections.
A few other small cracks were found, and a decision was made to replace the valves aboard Discovery with three that flew a dozen times previously on the same orbiter. They were even installed in the same positions.
In the meantime, additional tests were carried out to help engineers figure out what might have led to the crack and debris during Endeavour's launching. The damaged valve showed clear signs of high-cycle fatigue, which was a surprise to everyone. The environment in the pressurization line was thought to be relatively benign, without the sort of acoustics or vibrations that could lead to such stress.
Using computational fluid dynamics, however, engineers were able to identify heretofore unknown harmonic modes that could, in fact, produce the sort of forces that could buffet an extended valve and lead to the sort of fatigue seen in the cracked Endeavour valve.
At the same time, additional testing revealed small, undetected cracks in many valves that apparently were masked earlier by surface roughness in the valve material that were part of the manufacturing process. That meant the supposedly crack-free valves aboard Discovery were suspect.
That kicked off a major effort across multiple NASA field centers to find out what might happen if another crack led to another piece of debris being shot down the hydrogen pressurization line during launch. Mockups of the 79-foot-long pressurization line were built and small fragments similar to what came from the Endeavour valve were shot down the pipes at the same sort of speeds one might expect during a launch.
A possible long-term fix would be to polish off the surface roughness that can mask cracks, assuming additional analysis shows those changes would not significantly change the environment in the lines.
But the goal of the work this week was to show whether Discovery could be safely launched as is, even if one assumed a valve broke during flight. A punctured line would be potentially catastrophic. Initial test runs showed no outright punctures, but several areas of the line have "negative margins," sources said, and multiple variables are in play that make modeling the results of impacts difficult.
At the Stennis Space Center in Mississippi, engineers have conducted more than 60 test runs using a full-scale pressurization line flowing hydrogen gas to model the flow pattern, to determine how that affects the movement of debris in the line and whether that debris could cause a puncture.
Those tests are continuing, along with work to accurately model the pressurization line environment so computers can run detailed "what-if" scenarios to help characterize potential risk.
"The testing has been proving velocities and what the particle does as it travels down the line," Herring said. "Does it rotate, does it kind of fall into a set kind of movement? All of that is going very well, but there is just a tremendous amount of information that's being collected."
Some could view a decision to proceed with flight while working to fix a known problem as similar to what happened in late 2002 when NASA managers opted to keep flying while working to fix a problem with external tank foam insulation. The shuttle Columbia was destroyed during re-entry Feb. 1, 2003, because of a hole in the ship's left wing that was caused by impact with foam insulation from the external tank.
But others say the issue with Discovery is very different. NASA had a long history of foam impacts while the valve problem with Endeavour was the first in 124 missions. Even though the acoustic environment in the pressurization line is not what engineers expected, the valves currently aboard Discovery have no obvious flaws and a solid flight history.
If the testing this week and next shows debris impacts pose no significant additional risk, NASA managers may be able to clear the ship for flight. But there are major uncertainties at this point.
"Teams from multiple NASA centers and contractor sites have made significant progress in understanding what caused the damage to a flow control valve in shuttle Endeavour during its mission in November," NASA said in its statement. "The engineering teams have performed a tremendous amount of work, including computer modeling and actual tests to determine the consequences if a piece of a valve were to break off and strike shuttle and external fuel tank components. More time was needed to complete analyses and testing necessary to fly safely."