A paper given to the Windhover Society of the Oratory School on the 5th February 2003.
Dauphin, Ladies and Gentlemen,
Up until a couple of days ago, I was planning to present an entirely different paper to this Society. However, the events of last Saturday made me feel that perhaps the Columbia catastrophe might be a more appropriate subject than a little-known field of cosmological research.
For those of you who are perhaps not aware of what occurred, I will briefly summarize what the newspapers, radio or television should have told you.
At about quarter past one last Saturday afternoon, the 1st of February, the Space Shuttle Columbia fired its engines to begin the final stage of any space mission - re-entry. Three quarters of an hour later, sixteen minutes before landing, when the shuttle was 61 kilometres above the ground and travelling at 20 thousand kilometres per hour, contact was lost with the Shuttle and its seven crew.
To observers in Texas, immediately below Columbia's flight path when it was destroyed, it was all over in a matter of seconds. The shuttle was torn apart, littering a huge area around Dallas and Fort Worth in Texas with lumps of debris, the largest of which landed in a lake.
This mission - STS-107 - was a success story, right up to the bitter end. This was a mission dedicated to scientific research - to the extent that the crew was divided into a Blue Team and a Red Team, so as to be able to conduct experiments literally 24 hours a day, taking 12 hour shifts. They carried out about eighty experiments in a state-of-the-art Spacehab laboratory module carried in the cargo bay.
The experiments varied hugely, from investigations into the human metabolism's behaviour in microgravity through the growth of protein crystals to the evaluation of new stellar navigation techniques, and would have made this a very important mission even if everything had gone exactly according to plan.
So what actually happened in those last few minutes? From Mission Control's point of view there was no indication of anything wrong until 13:53 GMT, when flight controllers started not only getting unusual readings from some sensors in the left wing, but also losing some sensors completely. At approximately 14:00 GMT, after losing tyre pressure and temperature readings from the left wheels, Mission Control tried to contact the shuttle with the message, "Columbia, Houston, we read your tyre pressure messages and did not copy your last." The only reply before communication ceased was "Roger... erm...".
After several more futile attempts to regain contact with Columbia's crew, the Entry Flight Director put the contingency action plan into effect. The NASA Administrator, Sean O'Keefe, immediately notified President Bush as well as the US Secretary of Homeland Security that the shuttle had been lost. At the same time, NASA scrambled rescue units in a hopeless search for survivors. In fact, within half an hour of losing the spacecraft, NASA had already assembled a Mishap Investigation Team to investigate the loss. Because of the hazardous nature of the debris - due to a variety of toxic compounds and vapours - and also so as to speed the gathering of any useful evidence as to precisely what happpened, it is one of NASA's top priorities to gather the fragments as quickly as possible. However, the sheer size of the impact area - more than 2000 debris fields - means that it'll be a really long time before every bit of the orbiter is gathered. Let alone every bit of the crew - although some human fragments have already been found in several different locations near the Texas-Louisiana border. The recovery operation is being coordinated at Barksdale Air Force Base in Louisiana, assisted by the Federal Emergency Management Agency, the National Transportation Safety Board, the Federal Bureau of Investigation, and local police forces.
Over the last few days, the Mishap Response Team has been working pretty much non-stop to analyse the vast amount of telemetry data received from the shuttle during the last few minutes of its flight. The current consensus is that the disaster must have been due to damage to some of the many thousands of insulation tiles that protect the shuttle from the high temperatures generated by air resistance on re-entry.
The first anomaly in the telemetry, at 13:53 GMT, was the sudden loss of four temperature sensors connected to the Shuttle's left hand elevons. Columbia was flying 44 miles above California at about Mach 20. This was pretty much concurrent with an increase of about 35 degrees in bondline and strut temperatures in the left wing near the wheel well. An astronomer at the Californian Institute of Technology, Carmen Sanchez Contreras, reported seeing something falling off the shuttle at this time, but what it might have been is still unknown.
At 13:54 GMT, temperatures at the top of the bondline where the left wing joins the fuselage started to increase unusually, by 60 degrees in the following five minutes.
Four minutes later, at 13:58 when the Columbia was over New Mexico, telemetry showed a sudden increase in drag on the left hand side of the Shuttle, indicating structural damage to the wing, and an increase in left undercarriage tyre pressure. The data suggests that they didn't actually burst, however.
At 13:59 GMT, when the Shuttle was over western Texas, the autopilot logs show Columbia firing steering jets and trying to roll to its right to compensate for a further increase in drag on the left wing. Although the magnitude of the correction was well within the capability of the orbiter to carry out these maneuvers, the corrections made were larger than recorded during any other Shuttle re-entry. Simultaneously, data feeds from the undercarriage stopped being received.
Communications with the orbiter were lost a few seconds later at 14:00 GMT. At this time the Shuttle was flying 38 miles above eastern Texas, at the point in the re-entry trajectory where the shuttle experiences its maximum heat and stress. There's another 32 seconds of data, but it's scrambled and it'll probably be a long time before any sense can be made of it.
So, here's what happened, as I understand it. During re-entry, the shuttle is protected from extreme temperatures of up to 2000 degrees by a layer of 24 thousand tetrasilicide and borosilicate powdered glass tiles. Sometime during this mission, tiles on the underside of the left wing near the wheel bay were badly damaged, weakening their bonds with the orbiter's surface. When subjected to the stresses of re-entry, the damaged area of insulation tiles became detached - and once they came off, the ones next to them came off more easily, so the ones next to them came off, and so on right over the left wing. About six minutes before loss of communication, the rapid temperature increase in the area of the wheel well indicates that the heat shielding had just about all gone, and that the actual skin of the wing was starting to ablate - that is, the orbiter was starting to burn up. The increase in drag and in tyre pressures suggest that the the wheel bay door actually came off, increasing stresses on the already damaged left wing which were exacerbated by the autopilot corrections that soon followed. As far as I can tell, it's quite suprising that structural integrity was maintained for so long. Once the left wing finally parted company with the rest of the orbiter, Columbia - travelling at eighteen times the speed of sound - must have broken up in a matter of moments. The crew literally didn't stand a chance.
Columbia - Orbiter Vehicle 102, to give the spacecraft her official designation - was the oldest of NASA's fleet of five Space Shuttles, of which three remain. Commissioned in 1972, its first flight took place on 12th April 1981. In total, it flew 28 missions, including a highly successful flight to repair the Hubble Space Telescope last March.
Ablative tile damage has been noted before. On Columbia's first flight, STS-1, 16 tiles were lost and 148 were damaged thanks to an overpressure wave created by the Solid Rocket Boosters. There have only been a couple of rare missions when tiles have not been damaged by one thing or another. At the end of one Shuttle flight, a basketball sized hole was found burnt into one of the orbiter's wings where tiles had given way.
When the Columbia first flew, the ablative tiles were still under development, and for a while - one or two missions - the orbiter carried a tile repair kit, essentially a big syringe full of polyfiller. This was soon discontinued when it was realized that even if Mission Control (or the astronauts themselves) realized there was significant damage to the tiles, and knew whereabouts it was, it would be pretty much impossible for them to do an EVA to repair the damage. Most missions don't carry the special rucksack-like device that allows astronauts to maneuver around in microgravity. Equally, it's quite likely that in just inspecting or attempting to repair the fragile ablative surface the astronauts would do more harm to it than good.
In this case, the analysts think that they've found the cause of the critical damage. On January 17th, while reviewing videos of the launch, engineers spotted a lump of insulating foam falling from the giant tank that holds most of the fuel the Shuttle uses to get into space. The piece of insulation appeared to strike Columbia's left wing before falling into the engine exhaust and burning up. They immediately carried out an assessment of what sort of damage the foam could have caused, but concluded that any damage would be "inconsequential", posing negligible threat to flight operations or safety. This may or may not have been a justifiable conclusion. Firstly, the foam is very, very soft. It's like a very crumbly sponge. The fuel tank's orange appearance on the launch pad is caused by the couple of centimetres of this insulating foam applied all over to keep the liquid hydrogen and oxygen in the tank at the necessary cryogenic temperatures. Pieces of the foam frequently come away from the tank - the vibrations and temperatures involved at launch shake it loose. Pieces of it have been known to strike the cockpit windows before, but this stuff is so soft that the worst it does is to leave a bit of a smear. However, remember that the ablative tiles are also pretty soft. There are several instances in which damage to the tiles was attributed to pieces of foam falling from the fuel tank during launch, as happened on January 16th. However, engineering analysis has always indicated that the damage posed no risk to flight crew safety, merely increasing turnaround time between flights. On the other hand, this was the largest piece ever observed to fall and strike the shuttle.
Clearly the evidence for NASA being negligent is fairly ambivalent. You're probably aware that the media are currently making a big deal about Fischbeck and Paté-Cornell's 1994 paper, entitled "Risk Management for the Tiles of the Space Shuttle". In this paper it was recommended "That NASA... reinforce the insulation of the external systems (external tank and solid rocket boosters) that could damage... high risk tiles if it debonds on take off."
Taken at face value, this paper constitutes damning evidence. NASA was warned about exactly this accident occuring - 9 years ago - and didn't get the problem sorted out. In actual fact, the problem was addressed. By 1995, NASA had evidence that up to 90 per cent of damage to tiles was being caused by falling insulative foam. They believed that the problem was a process they called "popcorning" and initiated a research programme to find a solution. In 1999, the composition of the foam was modified, which was believed to have solved the problem. However, this was clearly not the case, as this latest incident demonstrates. Even so, NASA is not demonstrably at fault. They were informed of a risk, at took steps to fix it, and believed that the risk really had been sufficiently reduced. And with some justification - over the last four years, only one relatively tiny piece of the insulation debonding, with only superficial damage resulting.
Many other accusations in the media that NASA had ignored safety warnings seem to stem from the Aerospace Safety Advisory Panel's 2001 report, which stated that, "The panel's safety concerns have never been greater." However, these concerns were based on concerns about long-term planning to do with the replacement of the shuttle and its associated technologies, not to do with the safety of the shuttle at the present moment. None of the recommendations in the report have any bearing whatsoever on the circumstances of the Columbia accident.
Notably, the ASAP report also states that "NASA's operations evidence high levels of safety consciousness and sincere efforts to place safety foremost."
I find it very hard to take seriously accusations that NASA had ignored safety warnings - since the Challenger disaster, when NASA officials really were guilty of ignoring warnings about Solid Rocket Booster O-rings in cold weather, NASA has always been absolutely neurotic about safety.
Despite this latest incident, the Shuttle still has the best safety record of any spacecraft design. Given the hazards involved in space travel, it's very suprising that NASA has gone on for so long with so few casualties, no matter whether they heeded safety warnings or not. Consider a typical Space Shuttle mission. The Shuttle, when it's on the launch pad attached to the boosters and fuel tank, is one huge bomb. Take-off is a bit like setting the bomb off, but in a fairly controlled fashion. If a valve failed, or a pump jammed, or a tiny leak occurred, it's all over. Either the Shuttle would blow up on the spot like Challenger did, or crash shortly after launch.
Soon after the launch, there's the next risky phase of the mission - jettisoning the boosters and then, a few seconds later, the fuel tank. Because mechanical linkages can be unreliable, and the time of separation is critical, the components of the Shuttle system are linked by explosive bolts. Each bolt contains two charges, just in case one fails. On most shuttle launches, at least one if not more of these explosive charges fails to blow. If both the charges in just one of the bolts failed to blow, separation would not occur cleanly and, if the astronauts were lucky, the orbiter would be badly damaged. If they were unlucky, they would fail to make it to orbit, and would re-enter the atmosphere and burn up.
Once in orbit, the astronauts are in one of the most hostile environments known to man. If something goes wrong with the Shuttle that precludes it returning them to earth, there's currently no way to rescue them before they run out of life-support supplies. And the list of things that could go wrong is pretty long.
The return to earth is almost as hazardous as the launch. The Shuttle rips into the earth's atmosphere at collosal speeds, generating huge temperatures due to air resistance. If something goes wrong, you've got another Columbia. The Shuttle's an unpowered glider for the duration of the re-entry: so if the undercarriage fails to lower, or the pilot misses the runway the only option is to ditch in the sea - a very risky business, to say the least.
So you can see that given the risks it's actually quite suprising that NASA has gone on for so long with so few astronaut casualties. They can make sure components are reliable and backup-redundant as possible - but in the end, there's only so much they can do.
Where do we go from here? The International Space Station (or ISS) is the most expensive human artifact ever, having already cost more than 60 billion dollars to build, and it relies extensively on the Shuttle for resupply and construction flights. Although it can be kept going for a while by Russian Progress flights like the one that took place on Sunday, by June another Shuttle flight to the station must have taken place to relieve the current crew.
If the results of the current investigation indicate that modifications will have to be made to the existing Shuttle fleet, and if these modifications are projected to cost more than about four billion dollars, it's almost certain that a new spacecraft design using modern technologies would be more economical than to replace Columbia and upgrade the other orbiters. Bear in mind that the Shuttle design is thirty years old already. I don't even know anyone who's got a car that old.
Note that because the Space Shuttle does not include a crew escape system, it is below the standards that NASA is currently reviewing as the basis for future space vehicle designs.
Even if it is decided that the Shuttle programme is worth continuing, I think it likely that the Shuttle will be replaced within the next decade, most likely with a new design based on a resumption of the X-33 spaceplane project. Unfortunately, NASA's track record of encouraging new ideas isn't very good. They cut funding to the Delta-Clipper project, because they thought that it wouldn't work very well, then cut the X-33 because they thought that rockets weren't the way forward, investing instead in hypersonic research, while simultaneously slamming the innovative Rotol project as "gimmicky". Hypocrites.
Even so, in the not-to-far off future I expect to see a movement towards one-shot technologies like the European Space Agency's Ariane launcher both for satellite launches and manned science missions - even possibly for ISS resupply and crew relief flights. Although I haven't done a detailed comparison of Shuttle flight costs versus, for example, Russian Energia launches, I'm hoping to do so sometime in the near future.
I'd also expect the tile repair kit to make a comeback in some form or another.
It's important to remember that NASA's yearly budget is only 14 billion dollars, while the US Department of Defense gets 350 billion. And NASA's budget's being cut all the time, knock-on effects from which will significantly compromising mission safety. If you took what the US government spend each year on weapons of mass destruction and gave it to NASA, it would fund a new spaceplane design by itself. As frivolous as the space industry might seem to the uninformed, it is surely much more worthwhile to concentrate on exploration and colonization than on more efficient ways to kill people. As someone wrote to the sci.space.tech newsgroup on Sunday, "Make spacecraft, not war."
The most important thing, though, is that space research and development is not hampered by this specific tragedy. Those who died aboard the Columbia would not like to see the human race step back from space exploration because of their deaths - I'm certain they'd want us to see that going to space is very important to the human race's further development and, indeed, our survival as a species. We can't stay on this beautiful planet for ever. As Tsiolkovsky said, "Earth is the cradle of Mankind, but one cannot remain in the cradle forever."
Thank you for coming. I hope you've found this paper informative, and I'd be very happy to answer any questions you might have.