Shuttle launch delayed to at least Feb. 3 for additional tests, crack analysis

Editor's note...
  • Posted at 02:39 PM, 12/03/10: Shuttle launch delayed to at least Feb. 3 for additional tests, crack analysis
  • Updated at 07:25 PM, 12/03/10: Adding tanking test details
CBS News

KENNEDY SPACE CENTER, FL--Launch of the shuttle Discovery on a space station resupply mission will be delayed until at least Feb. 3, NASA managers announced Friday, to give engineers more time to carry out tests to help figure out what caused cracks in the ship's external tank and what, if any, modifications might be needed before the ship can be cleared for flight.

"It's time to pursue a different path, and that's to head out with some test data," Bill Gerstenmaier, chief of space operations at NASA headquarters, told reporters. "Basically what we're going to do with these tests is make sure we didn't overlook anything, we'll see if these tests can reveal any new information for us and it'll also help us sort out what the real problems are we need to be working on versus ones that we just think theoretically may be there."

Shuttle Program Manager John Shannon, explaining the need for additional tests to determine the root cause of cracks in the shuttle Discovery's external tank. The tests and additional analysis will delay launch until at least Feb. 3. (Photo: NASA TV)
Assuming the upcoming work gives NASA the confidence to proceed, a launch on Feb. 3 would be targeted for 1:34:28 a.m. EST, roughly the moment Earth's rotation carries launch pad 39A into the plane of the International Space Station's orbit. Landing back at the Kennedy Space Center would be expected the evening of Feb. 13.

If that schedule holds up, launch of the shuttle Endeavour on a mission to deliver a $2 billion physics experiment to the space station, which had been scheduled for Feb. 27, would slip to April 1 at 3:15:55 a.m. EDT. Launch of NASA's final planned shuttle mission, a station resupply flight with the shuttle Atlantis, would remain targeted for next summer.

But that schedule assumes engineers resolve major questions about the ability of Discovery's tank to withstand the rigors of fueling and launch.

Despite weeks of around-the-clock analysis, engineers have not been able to determine the root cause of four cracks in two adjacent structural ribs, or stringers, making up a compartment in the external tank that separates liquid oxygen and hydrogen tanks. The cracks were discovered after a Nov. 5 launch attempt was scrubbed because of a gaseous hydrogen leak where a vent line attaches to the side of the tank. That problem was traced to a quick-disconnect misalignment issue, repairs were made and NASA managers are confident the system will work properly the next time Discovery is fueled for launch.

Engineers also repaired the stringer cracks, splicing in replacement sections, attaching "doublers" to add strength and re-applying foam insulation. But figuring out what caused the cracks in the first place has turned out to be a major engineering challenge. Before Discovery can be cleared for launch, the community needs to understand the likelihood of new cracks forming that could lead to potentially dangerous foam debris during ascent or possibly even compromise the tank's structural integrity.

"We've hit a point where there is no obvious answer as to what occurred," said John Shannon, the shuttle program manager at the Johnson Space Center in Houston. "What that means is we have to take the next step, we have to look in greater detail to understand what types of stresses you could put in these stringers in the assembly process, see how they could line up to add stress to that stringer and we have to do that through a demonstration. Analysis is not going to get us there."

Discovery's external tank has been repaired, but an exhaustive analysis to prove the tank will stand up to the rigors of fueling and launch has forced NASA to delay launch until at least Feb. 3. (Photo: NASA)
Repeating a well-known phrase attributed to Wernher von Braun, Shannon said "one good test is equal to a thousand expert opinions, right? So we're at the point where we need that test, we need that fine level of data to understand exactly how those assembly stresses could line up to give us a crack when we initially loaded."

The lightweight aluminum-lithium alloy used in shuttle external tanks is known to be more brittle than the heavier aluminum material used in earlier fuel tanks. Cracks in aluminum-lithium components are not unusual -- an average of one per tank is found and repaired during manufacturing -- but Discovery's are the first to be detected at the launch pad, the apparent result of thermal stresses.

External tank hardware contracts or shrinks slightly when exposed to ultra-low-temperature rocket fuel. The cracks in this case occurred near the top of the intertank where the stringers meet a flange that runs around the circumference of the lower section of the liquid oxygen tank. Thermal stresses tend to pull the tops of the stringers inward as the oxygen tank is chilled, but the hardware is designed to accommodate that.

Engineers suspect some sort of unexpected built-in load was present that, in combination with normal thermal stress, resulted in stress-relief fractures in two stringers when the tank was loaded with propellants Nov. 5. But the precise mechanism is unclear and as a result, engineers are unable to accurately model the tank's response to the forces it will experience during another fueling and, eventually, launch.

To get a better understanding of what might be going on, two major tests are planned. One team of engineers will build a high-fidelity mockup of the ribbed intertank components that cracked, including deliberate flaws. The idea is to replicate, if possible, the damage seen during the Nov. 5 fueling.

At the Kennedy Space Center, strain gauges and temperature sensors will be attached to Discovery's tank before a fueling test later this month to collect data on the actual thermal stresses experienced when liquid hydrogen and oxygen are pumped aboard. Along with the sensors, stereo cameras will be used to measure exactly how much the tank shrinks during fueling.

"We're hoping the optical piece, plus the strain gauge measurements, will give us a really great indication of what the stress level is in those stringers," Shannon said. "We believe the design is robust and should not fracture under that stress. We'll verify that and then we'll add the assembly issues you could potentially have to see if we can get to a root cause.

"We were hopeful early on that it would be some kind of very obvious kind of flaw. Didn't happen. Then we were hopeful that just a simple cryo tanking would cover us for any ascent loads. It's very close, but it's not quite there. So now we have to go to that next level ... to get the root cause and determine what our screening criteria is to fly that tank confidently."

Engineers hope to carry out the test by Dec. 16. That would allow enough time to haul Discovery back to the Vehicle Assembly Building for additional X-ray inspections, if necessary, and still get the ship back out to the launch pad in time to meet the Feb. 3 target date. As of this writing, there are no plans for a roll back. But if NASA managers ultimately decide to inspect stringers on the opposite side of the tank from the shuttle, a roll back would be required because of access problems at the pad.

Whenever it occurs, the fueling test will mimic an actual launch countdown.

"I want to do a test in a flight configuration," Shannon said. "What that means is that we would remove foam, that nice foam they just put on the tank (for repairs), we would go remove that, we would put our instrumentation on there and then we're going to foam it back up. Because I want the most accurate models I could possibly get of the stress and the temperatures in that area in a flight configuration. That takes time. That's what's kind of driven us out of the December window."

NASA originally hoped to launch Discovery Nov. 1, but the flight was delayed one and then two days because of work to replace quick-disconnect fittings in the shuttle's right-side orbital maneuvering system rocket pod. Launch then was delayed 24 hours to Nov. 4 because of an electrical glitch in the circuitry associated with a main engine controller. Troubleshooting showed the likely cause was transient contamination in a cockpit circuit breaker.

A launch attempt Nov. 4 was called off before the start of fueling because of stormy weather at the Kennedy Space Center. The launch attempt on Nov. 5 was cancelled because of the gaseous hydrogen leak. After the cracks were discovered, launch was delayed to no earlier than Nov. 30. Launch then slipped to no earlier than Dec. 3 and finally to NET Dec. 17. It became obvious after a PRCB meeting Thursday that more time was needed to resolve the problem and launch was delayed to at least Feb. 3.