Laying out the shuttle on the grid made it easy to see patterns—burns, cracks, stress marks, and the like—in the pieces of the orbiter structure. Those could hint at the state of the vehicle when it broke up. Items from the external surfaces were to be placed on the grid. All materials from the crew module—the compartment in which the astronauts were sitting for launch and landing—would be reconstructed in three dimensions inside a walled-off area with controlled access. Debris from the shuttle’s internal systems—wiring, plumbing, avionics, and so forth, as well as materials from the modules in the payload bay—would go onto “bread racks” along the walls of the hangar. These items were thought not to have contributed to the accident, so they were segregated from the structural debris but readily accessible if necessary. Also lining the hangar walls were workstations for the various engineering disciplines to examine pieces of debris as they arrived from Barksdale.
We set up a separate decontamination site outside of the hangar to protect its workers from potentially hazardous materials. It was possible that some debris might arrive from Texas still contaminated with hypergolic fuels. Pyrotechnic charges might also be attached to some of the wreckage, and these would need to be handled off-site. There was a remote possibility of exposure to biological hazards, either from experiments on the Spacehab module or even from human remains that might have inadvertently made their way back to Florida in the debris. Fortunately, this did not turn out to be an issue, but we needed to have a process established for any contingency.
The NTSB suggested the personnel receiving the debris make a rough hand sketch of the appearance of every item as it arrived. Technicians could also write notes to be filed with the sketches. “That process quickly went out the window, as soon as the trucks started arriving,” Altemus said. Neither NASA nor the NTSB had anticipated we’d recover eighty-four thousand pieces of debris—and most of them would be the size of a nickel. Sketching every item was simply too much work. However, we did photograph and bar code every item received—a level of detail far beyond that in typical NTSB investigations.
With the Columbia Accident Investigation Board’s concurrence, we eventually opened up a second location in an equipment hangar adjacent to the Apollo Saturn V Visitors Center across the road from the Shuttle Landing Facility. This building held friable materials that posed a health hazard because of their microscopic fibers. Some of the items placed in this building were Kevlar-wrapped pressurized tanks and pieces of the payload bay doors.
—
KSC engineers and technicians were eager to work in the hangar. The people who had tended Columbia over the years desperately wanted to help out with the accident investigation in some way. Many of them had volunteered to be deployed to East Texas to help in the search, but for one reason or another, their management had told them that it might not be the best use of their talents. Despondent over the loss of the shuttle, they were frustrated about the lack of opportunities to do something “useful” in the aftermath to take their minds off their grief. Working in the reconstruction hangar was not going to be an easy task. However, it would be a welcome chance for people with specialized knowledge of the shuttle’s systems to use their talents to solve the mystery of what doomed Columbia and her crew.
Orbiter Project Office head Ralph Roe asked Jon Cowart to be the NASA engineering lead from Houston for the reconstruction. Warren “Woody” Woodworth was designated as the chief Columbia reconstruction engineer for United Space Alliance (USA). Jim Comer helped Steve Altemus set up the hangar and then became the head of operations at the hangar for USA, overseeing about three hundred contractors in the reconstruction.
The Astronaut Office assigned Pam Melroy to manage the crew module reconstruction, because she was the lead “Cape Crusader”—the astronauts assigned to work on shuttle issues at KSC. Melroy and her colleague Marsha Ivins alternated workweeks in the hangar, until Ivins moved on to other duties about a month after the accident. Melroy and her staff of six people would have the unenviable task of examining what was left of the crew module and the equipment used by the astronauts. Two of her key assistants were John Biegert and Robert Hanley.
Biegert, a long-term member of the flight crew systems engineering group from Houston, supervised the teams that loaded all of the crew equipment, lockers, food, EVA suits, experiments, and other items—five thousand to seven thousand pieces of equipment and materials altogether—into the crew module in the ten days prior to Columbia’s launch. He and his teams knew where every item should have been stowed when Columbia’s crew prepared the ship for landing.
Robert Hanley—one of the last people to see Columbia in flight—had been going “stir-crazy” in Houston and was desperately looking for something useful to do. He asked permission to visit KSC and see the reconstruction effort. He only intended to be there part of the day and return on the afternoon flight back to Houston. He walked into the crew module area and saw Pam Melroy writing on a whiteboard. Assuming that he had come to help identify debris, she said, “Robert! Great! How long can you stay?” He replied that he just needed to go back to Houston and get some clothes, and then he could stay as long as she needed his help. He became Melroy’s lead engineer.
Renée Ross was an example of someone brought in for specialized expertise. Ross, a USA employee who managed Columbia’s flight-data books, received an email from Melroy, asking her help in identifying some fragments of pages that appeared to be from various manuals in the crew module. After a few weeks of exchanging photos of documents via email, Melroy asked Ross to come to KSC to work through the accumulated backlog of crew module document remnants that were being returned from the field. Ross eventually identified and cataloged about four hundred pieces of documentation from Columbia.
Thermal protection system lead engineer Ann Micklos, who had been dating Columbia crewman Dave Brown, received a call from her chief engineer. Micklos recalled, “Knowing my circumstances, he told me, ‘We really need you to be part of the reconstruction crew. You know Columbia’s systems and you’re the right person to be there. Is this something you think you can do?’ Well, it’s human nature to want to help during a time like this. There was no way I could just sit in my office. Obviously, there was no place I’d rather be than with Columbia, even with how difficult the challenge initially was.”
Finally, some of the “founding fathers”—the original designers of the shuttle at Rockwell International, now a heritage company within Boeing—came to the hangar to help. John Tribe and Sam Kreidel were among the senior Rockwell engineers who had lived with the shuttle since 1972. They knew how the shuttle was originally designed and built.
In contrast, the engineers and technicians from Kennedy knew the current status of the systems and components some twenty-odd years later. There were no digital drawings of most of the shuttle’s components, so the combined knowledge of the designers and the hands-on staff was needed to identify some pieces of debris.2 For example, the outboard tires on the left main landing gear could only be specifically identified by the presence of balancing patches that were installed at KSC.
All told, roughly four hundred scientists, engineers, and technicians worked in the hangar during Columbia’s reconstruction. Most of the workers were from United Space Alliance (who maintained the shuttles at KSC) and Boeing. Nearly every NASA center was represented.3 Operations ran in two shifts per day, six days per week.
—
Several of the people who reconstructed Columbia had also worked with Challenger’s wreckage following the 1986 accident. Even though the circumstances of the accidents and the reconstruction processes were vastly different, some lessons learned from seventeen years earlier assisted in the Columbia reconstruction.
Challenger fell into the ocean without having left the atmosphere on her mission, traveling a little less than twice the speed of sound at peak velocity during ascent. Despite the way the accident appeared on television, the shuttle stack did not explode. Rather, the vehicle broke up becau
se of aerodynamic forces after the external tank structure failed. Most of Challenger’s debris—much of it in large chunks and sections—came down in a relatively confined area within sight of the Florida coast.
Columbia, on the other hand, had been traveling in excess of Mach 18 at an altitude of over two hundred thousand feet when it disintegrated. Its wreckage was twisted, shredded, subjected to plasma, melted, oxidized, burned, and scattered over a 250-mile-long path. The vast majority of the debris that came back from Columbia was smaller than an office desk. Much of it was the size of a nickel.4
NASA’s primary goals in the Challenger recovery had been to retrieve the crew, the right-hand solid rocket booster (SRB) with the O-ring that had burned through and caused the accident, and the crew module. NASA also was interested in debris from the shuttle’s payloads, the left-hand SRB, the external fuel tank, and a few other specific components. The accident investigation was not focused on the orbiter itself, because the vehicle was clearly not the cause of the accident. There was no need to document the latitude and longitude of where each item was found.
Navy ships and divers spent seven months scouring the sea floor for the priority pieces of Challenger’s debris. They retrieved 50 percent of the SRBs and less than 50 percent of the orbiter and external tank.5 About 70 percent of the crew module and the surrounding fuselage structure was recovered.6 The remains of Challenger’s astronauts were inside the crew module when it was recovered. The crew compartment had been in the water for more than a month before the navy located it.
Some parts of Challenger unrelated to the accident were left on the seafloor rather than incurring the risk and expense of retrieving them.7 Pieces of Challenger’s wreckage occasionally wash up on the Florida beach to this day, decades after the accident.
NASA reconstructed parts of Challenger in KSC’s Logistics Facility, about one mile south of the VAB on Contractor Road. I clearly remember it. I was working in design engineering and launchpad safety and structural systems at the time, and I was one of the employees permitted to see the debris during Challenger’s reconstruction. My first impression was the horrible smell. Challenger reeked of seawater and rotting barnacles. A few of the shuttle stack’s potential failure points were laid out on the floor, including the SRB aft field joint that burned through and the attach point from the SRB to the external tank. A few sizable pieces of the fuselage were propped up using two-by-fours in a rudimentary step toward reconstructing the ship. Much of the smallest debris was basically swept into piles off to one side. The crew module debris reconstruction took place in a small building adjacent to the Logistics Facility, with tightly controlled access.
John Biegert worked as part of the reconstruction teams for both Challenger and Columbia. He said that perhaps the biggest difference he saw between the two efforts was due to the advancement in digital technology over the course of seventeen years. Instead of using Polaroid cameras and the very limited computer technology of 1986 to document the debris, the Columbia accident investigation benefitted from digital photography, three-dimensional scanning and modeling, sophisticated computer databases, and other relatively new technologies.
—
The Columbia Accident Investigation Board wanted to maintain an on-site presence at the reconstruction hangar. I worked with them to determine where their offices would be located and how they would interact with the debris and the workers at the hangar. The CAIB set up in office trailers parked outside the hangar. This enabled them to observe what was going on, provided them private workspaces, and kept the Board from distracting our engineers and analysts who were trying to do their own work. This was not done to protect the information—it was to protect our people from having too many cooks in the kitchen.
We had an interesting relationship. If the CAIB wanted something, we were obligated to give it to them—but they could not direct our team. They could make requests, and we would fulfill the requests. The process seemed slow to them, but we did not drag our feet. We were doing it at the pace we were used to at KSC, which was meant to be methodical.
Greg Kovacs, a medical doctor and professor of electrical engineering at Stanford, came in toward the end of February as an advisor to the CAIB. He spent almost all of his time in the hangar with us, and he became a good friend and de facto member of our team. He jokingly observed that the “NASA approach to things is to put 50,000 people in a line and move forward an inch at a time.”
It did seem painstaking and slow sometimes, but we were following a basic tenet of the NTSB’s investigation procedures—to avoid speculating about the cause of an accident for as long as possible. The NTSB’s long experience proved that someone whose mind is latched onto a given theory will pursue that line of investigation and disregard evidence that points to other possibilities. That could lead to dead ends and wasted time. In the long run, the NTSB said, keeping an open mind as long as possible would actually speed up the investigation.
We learned and recited the NTSB mantra: “Keep an open mind. Let the debris tell the story. Collect the debris, and the debris will tell you what happened.”
In the end, the investigation concentrated on the problem areas identified in the fault tree. We knew the problem did not start on the right wing, so we didn’t spend a lot of time in detailed examination of the right wing. We knew where we had to concentrate our efforts, but being able to compare Columbia’s right wing to its left wing was invaluable throughout the process.
—
The NTSB was initially concerned that our workers might be overwhelmed in trying to deal with the quantity of wreckage arriving from the field. However, this turned out not to be an issue. Because Steve Altemus set up and staffed the hangar and its receiving processes so quickly after the accident, the facility and its personnel were ready to go as soon as the first two truckloads of debris left Barksdale on February 12 and arrived at KSC on February 13.
Barksdale dispatched a shipment to us whenever enough material filled a flatbed or an enclosed eighteen-wheeler. In the first several weeks, shipments arrived at KSC every other day. Just one month after the arrival of the first truckloads of material from Barksdale, the hangar already contained 33,798 pieces of debris, totaling 43,200 pounds, and representing more than 19 percent of the shuttle’s dry weight. The frequency of shipments began to tail off slowly as recovery operations cleaned out the debris field—first to a couple of shipments per week, and finally only one truckload per week.
KSC security cars escorted the shipments, and NASA special agents alerted state police along the route that the trucks were coming through. All of the NASA centers asked to participate in the honor of escorting Columbia back to KSC. To share that solemn role among the centers, one special agent from KSC and one from one of the other centers usually staffed the escort vehicles. These special agents also personally carried any sensitive materials being sent to KSC, rather than leaving them in the back of a truck. It was an eighteen-hour drive from Barksdale, with stops only for food and fuel.
KSC security special agent Linda Rhode (whom the reconstruction team nicknamed “Agent 99”) accompanied one shipment back from Barksdale with her boss, Mark Borsi. “It was four o’clock in the morning, and it was pitch-black. I was driving, and we’d been spending probably fifteen hours staring at the back end of this truck,” Rhode said. “Somewhere along Interstate 10 in the Florida Panhandle, a law enforcement vehicle jumps between me and the truck and pulls it over.” Borsi got out of the car and told Rhode to stay put. She said, “The agricultural inspector was talking to the truckers and then got a radio call. I think he realized the error of his ways. He hopped in his cruiser and took off.”
“Truck day” was always a special day in the hangar. It was not quite a celebration, but every load of new material meant that we were getting closer to figuring out what happened to Columbia. The staff and I went out to greet the drivers and thank them for bringing the material to KSC. Drivers told us where the material had come from. For some personnel
in the hangar, this was their primary news source about events in the field.
As the truck was unloaded, every item was “sniffed” to ensure that it was not contaminated by hypergolic propellants. Staff was on the lookout for crushed tiles and other friable items whose fibrous materials could cause respiratory problems.8 The processing centers in East Texas did their jobs well, because none of the material was contaminated by the time it arrived at KSC.
Materials were then triaged just inside the hangar from the parking lot. Every item arrived bagged (if it was small enough) or wrapped in cling-wrap plastic film, and carried a tag with an identifying number and the GPS coordinates where the piece was found. Quality assurance staff bar-coded, photographed, and cataloged each item into the database. Items that were related to payloads, fuselage, or internal structural items went to the appropriate engineering stations. Things that were obviously crew-related and debris from the crew module went directly to the crew module room. Other debris was examined to try to determine where on the shuttle it might have come from. Then it went into the bread racks or onto the floor grid as appropriate.
Identifying debris that may or may not have been from the shuttle proved an interesting challenge. One landing-gear strut was heavily oxidized and caked with mud when it arrived. “Before we got all that mud and crap off it, it looked like an anchor,” said Jim Comer. After cleaning it up, many people believed it was from a B-52 bomber that crashed somewhere in Louisiana or East Texas many years ago. On February 14, technicians pulled the endcap off to look for an inspection stamp inside the strut. Only then did we prove the piece to be from Columbia.9
Bringing Columbia Home Page 22
