Getting to Liftoff
NASA's mission management team puts a new philosophy on the flight line for Discovery.
Wayne Hale's nightmare came true on a midsummer Wednesday last year in Houston. Then deputy manager of NASA's space shuttle program, he didn't notice the time exactly, but he knew it was pretty late. After all, the stars were out. And a shuttle was up there among them-finally. It was July 27, 2005, 30 difficult months after the orbiter Columbia disintegrated in a re-entry catastrophe that claimed seven astronauts' lives.
Now it was showtime for Discovery, which had launched the previous morning and completed a couple dozen orbits of Earth. Seven astronauts soon would awaken to their third day on board, one featuring a tricky rendezvous with the international space station and the first shuttle visit to the U.S.-led outpost in almost three years.
Outside Mission Control, Hale had revved up the old Suburban and pointed its headlights toward home. NASA's most experienced flight manager was alone with his thoughts: two-and-a-half years . . . one-and-a-half billion dollars . . . hundreds of engineering tests, designs and redesigns . . . calculations, recalculations . . . analysis . . . work, rework . . . brought in the best minds . . . best technical thinking . . . reduced the potential for fuel tank insulation to fly off and strike the orbiter . . . convinced ourselves . . . Hale's phone jangled. It was the shuttle engineering and integration chief, John Muratore. "Wayne, we've got this imagery that, when you enhance it, you can see the foam hit the wing," he said, and Hale just about drove off the road. It was the lowest of the low points he'd seen in his 27 years as a flight controller, with the exception of the two shuttle accidents.
What a roller coaster of emotion the past 36 hours had been for Hale, the 27 senior government and contractor engineers on his mission management team, and some 500 technical specialists. From euphoric heights upon Discovery's launch, they sank to the depths of despair soon after liftoff. Video replays proved that NASA and its external fuel tank contractor, Lockheed Martin Corp., had failed in their efforts to comply with the foremost recommendation of the Columbia Accident Investigation Board.
A tiny TV camera mounted on the 154-foot propellant silo told the graphic truth: Between 129 and 149 seconds into Discovery's climb to orbit, several chunks of insulating foam shook loose from the tank. The biggest debris was a yard long, a foot wide, and about half the mass of a 1.67-pound chunk that struck and doomed Columbia in 2003. The accident investigators told NASA to eliminate debris shedding entirely to spare the remaining three orbiters a similar fate. The composition of the foam made that an impractical challenge, so NASA worked hard to guarantee that nothing heavier than about half an ounce would break free.
"We were wrong," Hale's boss told the news media that afternoon. The strained admission from then-space shuttle program manager Bill Parsons (now deputy director at Kennedy Space Center) hung like icicles in the air-conditioned auditorium at Johnson Space Center. "We have more work to do."
Hale, an avuncular, church-going guy, had a penchant for philosophical pep talks distributed in memo form. But in an e-mail to space, he got right to the point. "I'm absolutely mortified," Hale told Discovery's crew. "And we're not going to fly again until we fix it."
The foam hit was caught by a Sony XC-999 "lipstick cam" mounted on the fuel tank's exterior, near two struts where the orbiter is attached. Analysts had missed it in the live transmission beamed to the ground at 30 frames per second during the first two or three minutes of the shuttle's climb. Twenty seconds after the biggest chunk of insulation got loose, three smaller pieces popped off a nearby spot on the fuel tank. One small piece seemed to ricochet off Discovery's starboard wing. Columbia's hit also was on a wing.
Discovery was the most photographed shuttle so far, thanks to a beefed-up suite of imagers NASA employed to make sure it could obtain three useful views of the launch-another accident board requirement. The agency increased still, video and motion picture film camera coverage by 25 percent. A total of 113 government cameras were pointed at Discovery during ascent, including six on the shuttle. To look for falling debris, NASA added radars several miles north, south and east of the launchpad and positioned two high-flying WB-57 research planes with color and infrared cameras in nose-mounted ball turrets.
Talk about drinking from a fire hose. Mission managers had more information about the shuttle's physical condition than ever, and they had a deadline for getting it analyzed. Three days was the plan. On Day 4, the astronauts would use a laser camera system attached to Discovery's robot arm to make focused, 3-D inspections of problem areas that showed up in the analysis. That would leave roughly 10 days to try repairs and to come up with other remedies, if necessary, before the shuttle came home.
Getting the Picture
The Columbia Accident Investigation Board had drawn a bull's-eye around the mission management team (MMT), blaming it for "missed opportunities" to gather spy satellite pictures and other information it might have used to save Columbia's crew. The board prescribed a cultural makeover for NASA. For Discovery, Hale put his team at the point of the dart, requiring the members to meet regularly to deal with cooked-up crises under the tutelage of organizational behavior experts. Occasionally he integrated the simulations, synching up about a dozen NASA field installations, federal laboratories and contractor offices from California to Canada to Connecticut in a closed-circuit teleconference.
They met like this for several hours every afternoon during the mission. The daily MMT was supposed to demonstrate a cultural shift in the same way the image assets were supposed to verify the technical changes. Hale directed the proceedings from a networked command center he personally designed to invite discussion and accommodate dissent. The table is round-no head, no power center-and nobody is permitted to sit in the same place twice. The 1 o'clock sessions were all about communicating the results of analyses mostly done while the mission mana-gers were sleeping.
Coordinating a support team numbering at least 200 to analyze imaging at three field installations was Muratore's responsibility. When everything was being put in place ahead of the first big sim in early 2004, he realized he needed help. "We weren't ready," he says. "We didn't see how we were going to keep the activity in all three imagery labs and the inspection teams coordinated."
Terri Murphy's timing couldn't have been better. She walked into Muratore's office one day to deliver a report on the laser boom and walked out with the brand-new title of imagery integration manager. The 40-something computer scientist was an old hand at flight operations. Early in her career, she wrote code for Mission Control's data processing systems. She controlled onboard computers from the center for 10 years before the Columbia mission. Image analysis wasn't anything close to a real-time operation in the past, but that's what would be needed for the Discovery mission. The plan for converting visual inspection data into useful information for the mission management team was driven by the clock. Murphy sorted out how the assets would be used, how the data would flow between the labs and how their analyses would feed into on-orbit surveys for damage.
Months later, she remembers Discovery as the most rewarding assignment of her career. At the time, the traditional friction between NASA's operations and engineering divisions made it incredibly frustrating. The two mix about as easily as oil and water. They're always at odds about whether an answer is good enough. Operators need answers on time. Engineers like to answer questions whenever they're done studying them. Murphy often heard the protest, "We'll just continue to do it the way we've always done it. It's always worked in the past." Even after what happened to Columbia, she says, "I don't think that everybody recognized that maybe it didn't always work in the past."
Some thought the simulations were a nuisance. Others insisted on prolonging intercenter rivalries dating back decades. A few old-timers made it clear they weren't interested in taking directions from a woman. Her elaborate imaging network had a couple of holes she couldn't plug in time for Discovery's launch, but that's why there were three labs. If one lab didn't want to play, the other two could take up the slack.
Finding a Fix
A big loss of foam and something hitting the wing were the two "Oh, shit!" events Murphy was looking for. To her relief, a closer check of the time-coded images showed the debris shedding at high altitudes where aerodynamic forces are negligible. It was hard to say whether the small piece hit, but if it did, it didn't do serious damage. Poring over the images with Hale and Muratore in the Johnson lab that night in July, Murphy wondered to herself, "What else are we missing?"
She found out in a matter of hours. As Discovery closed within 600 feet of the space station, Commander Eileen Collins put the orbiter through an unprecedented perilous maneuver-a 360-degree back flip to show its nose and belly to residents on the outpost. They pressed 400 and 800 mm lenses against two windows and snapped away while the shuttle slowly spun. The rendezvous photography revealed two problems. Two spacers were sticking out of Discovery's tile heat shield like whiskers on a chin. Up by the commander's windshield, on the left side of the nose, a thermal blanket was torn and puffed out.
Hale's first inclination was to dismiss them as mere maintenance issues, just stuff to fix after Discovery returned to Earth. He says most of the mission management team felt the same way. And he acknowledges without being prompted that is the same red flag raised by the Columbia investigation. But this is the new NASA, Hale says. In the old days, "We would have relied a lot more on . . . wishful thinking."
Environmentally speaking, space is an uninviting place. Temperatures swing from 200 degrees Fahrenheit to minus 200 every 90 minutes as the shuttle travels in and out of Earth's shadow at a speed of 17,500 mph. Coming home from a visit to space is extremely unpleasant. Friction ignites a 3,000-degree inferno around the shuttle when it plunges into the planet's ever-thickening atmosphere. Inflate a bicycle tire with a hand pump. The pump gets hot as it pushes air. Reentry is the same concept, faster and hotter.
To protect its aluminum airframe and critical systems from heating extremes, the orbiter is covered with glass. The black belly is an intricate mosaic of 23,000 featherweight silica tiles so strong they could be dropped into cold water immediately after baking in a 2,200-degree oven without breaking. The upper fuselage consists of 2,300 white silica fiber quilts, each one custom sewn by a high-tech tailor. Thermal protection system technicians go to great lengths to make sure the heat shield is smooth. No tile can rest more than one-tenth of an inch taller than another, and gaps wider than two-tenths of inch must be filled with strips of ceramic fiber. Measurements are checked with lasers.
The tiles tend to chatter on the way uphill, so gap fillers shaking loose was no surprise. Technicians have noted hundreds of charred or missing in post-flight inspections over the years, but the first time they counted them in space was during the Discovery mission. As soon as Chuck Campbell saw two protruding, he put his 100-strong orbiter entry aero heating panel in crisis mode. One of NASA's most knowledgeable aerodynamicists, Campbell knew the history of this problem.
A rough surface disturbs air flow on the shuttle's underside. The resulting turbulence can make the surface hotter and the shuttle trickier to fly. The air churns itself up naturally when the shuttle descends. It happens at eight or 10 times the speed of sound if all goes well. Say something makes it happen earlier. Gap fillers have done this at Mach 18 twice in 25 years. Two shuttles landed with scorched and melted tiles in need of repair.
Discovery's spacers stuck out as much as an inch. Campbell's team couldn't guarantee the heat-up wouldn't happen too early and trigger another re-entry disaster. The mission management team had a lot to learn. Campbell told them, "Anything we do beyond Mach 18 is extrapolation or, at best, applying a theoretical understanding to something that you don't have any real data on." It took three days of computational analysis and a lot of explaining to make the managers understand the difference between comfort and conjecture.
Six days into the mission, the MMT decided an astronaut should try to remove the gap fillers during a spacewalk. Until now, the orbiter's belly had been a forbidden zone. It wasn't visible from the cockpit-a safety concern-and folks believed the myth that the tiles were too fragile to touch. A team of 200 contingency specialists back on Earth came up with the plan. On his ninth day in space, the flight engineer would ride the robot arm to the site, talking all the way. If he couldn't be seen, at least he'd be heard. The pressure-suited astronaut would pull the gap fillers out or cut them off. He kludged together a hacksaw with duct tape and a blade from the cockpit just in case. In the end, the only tools he needed were his gloved thumb and forefinger. Crisis averted.
Not so fast. What about that loose blanket?
A thermal analysis already showed that Discovery could land without any heating concerns even if the blanket came off. "But if it comes off, where will it go?" Justin Kerr piped up with that question on Day 6 when the mission management team was too busy with gap fillers to pay much attention. Kerr led the foam impact tests that pinpointed the cause of the Columbia disaster. He led the debris transport analysis team for the Discovery mission.
He and 80 colleagues burned the midnight oil computing an answer to his question and shooting pieces of blanket at samples of tile to see what would happen. Two days later, the 35-year-old put on a crisp white shirt and a striped tie and went to brief a group of noticeably agitated senior officials about the dire possibilities. They didn't think the blanket would come off, but if it did, and Discovery was traveling Mach 6 or slower, it was big enough-20 inches long, 4 inches wide-to cause grievous harm to the shuttle's tail and one of its two maneuvering engines.
The options were few: leave it alone, fix it in a spacewalk, or tell the crew to abandon ship and take shelter on the space station until they could be rescued. The safe haven concept meant risking another shuttle and crew. Spacewalking was an ugly option, too. The blanket batting is nothing but needles of glass. When the contingency specialists tried working with it, the fibers went everywhere. If they were uncontrollable in normal gravity, they could wreak havoc in zero-G.
Hale had an idea. Take a blanket, put it in a wind tunnel, and "buffet the wazoo out of it," he told the team. "If we'd kicked [it] off days ago, we probably could have accomplished [it]," Kerr replied. It was Aug. 2. Discovery was due to land in six days.
"You have posed a mortal threat to the vehicle," Hale said, leaning forward on the conference table and looking over the top of his thick eyeglasses. "People should not sleep, or take days off, till we get an answer to this problem. Is that clear?"
It takes weeks, more often months, to plan and execute even a simple test in a wind tunnel. Doing it in 48 hours is unheard of. But that is what Kerr's team did. They yanked blankets off a sister ship at Kennedy Space Center in Florida and built three test articles. They got stereo images of Discovery's blanket and damaged the test articles until they looked alike. They jetted them across the continent in a NASA Gulfstream 2 on Aug. 3. While the plane was in the air, mission managers rejected the test plan. It raised more questions than it answered. They revised the plan within an hour, before the plane landed at Ames Research Center in California. They completed a confidence-building run in Ames' Mach 6 wind tunnel about an hour before the management team convened on Aug. 4. The test articles remained intact.
Discovery landed without incident at Edwards Air Force Base in California on Aug. 9, after a one-day weather delay. NASA gave one of the gap fillers to the Smithsonian Institution. The damaged blanket and the identical test article are being mounted on a plaque to hang in Mission Control. Wayne Hale now manages the space shuttle program. He won't have a seat at the mission management team table when the next shuttle flies, but he'll be at the ready, like hundreds of others, with sage advice and support.