SpaceRef

SpaceRef


Transcript: Columbia Accident Investigation Board Roundtable June 12, 2003 (part 2)

Status Report From: CAIB
Posted: Friday, June 20, 2003

MS. LAURA BROWN: Okay. Mike?

MR. MICHAEL CABBAGE: Mike Cabbage with the Orlando Sentinel.

What sort of analytic work have you done to see whether the cracks that you've seen in the testing so far on panel six could propagate into the type of damage that eventually caused Columbia to break up? I can't - from what you know, can cracks of small and displacement of the type you've seen, lead to the kind of break-up that we saw?

MR. HUBBARD: The only thing that I've got at the moment is, you know, gut feelings from engineers that helped design this. There is no analytic work done to date. There is a program called “FLAGRO” - F-L-A-G-R-O - which was used on the shuttle liner crack and it took a million hours of super computing time to get that one so it duplicated what was seen. I have no doubt that people will be running these kinds of programs, but they take a while. So, we don't have any analytic evidence right now.

And as I said at the time, this is why you do the experiment. We're learning as we go.

MR. CABBAGE: Well, if I could just follow-up to that. So, there is - you really have no evidence at all at this point then, that those cracks could have propagated into the sort of damage that we saw go through the left wing on the sensor data and the remnants and what's your gut feeling on that?

MR. HUBBARD: We don't have any analysis done. Materials scientists know that cracks propagate - stresses cause cracks to propagate. You know, that could be temperature effects, it can be mechanical effects. I gave you a little list of the sort of things that this crack might have been - may have experienced.

On the other hand, the - you know, some gut intuition of people says, well, that probably would be fairly stable. The answer to that question is unknown and that's why we're continuing to do the work to prepare for the panel eight, nine, ten shot which is much more representative of the - before the breach occurred, at least that's where we think it occurred from all the other evidence and that's why we're collecting some more data.

MS. LAURA BROWN: Okay. Todd?

MR. TODD HALVORSON: Todd Halvorson of Florida Today.

I was wondering what your thoughts are on the differences between the two materials you are using: fiberglass and reinforced carbon-carbon. The pretest predictions on each of those materials and then what you've actually seen in the two tests to date.

MR. HUBBARD: Let's talk about the most recent tests and there what we observed was that pretest predictions on RCC and what actually happened were pretty good in terms of total load and total stress. The surprise was that we didn't get a break in an area with very high stress - three times the supposed limit. We did get a break in a place that was just, you know, barely above the limit. Now, what does this tell you? I think one of the earlier conclusions is this is not a simple tension or compression event. It may have elements of torsion. It may have, you know, where something is bending and twisting at the same time and that is kind of the feeling that the structural engineers are thinking about. And so, they're going to try to take the simple models that they started with and put in this torsional element. You know, the fact that it may not just be a bending, but maybe a bending and a twisting simultaneously. That could account for what we've seen.

The fiberglass, we know that it is a different material. It was chosen years ago as being a fairly good surrogate for RCC. I mean the panels we got, we didn't manufacture. They came off of Enterprise. And we thus far have done reasonably well in extrapolating from one to the other. As long as we stick to kind of a qualitative story, I think we'll be fine.

MS. LAURA BROWN: Okay. Gina?

MS. GINA TREADGOLD: Gina Treadgold with ABC News.

Is there anything more on the Day Two Object? Since that crack was sort of halfway down the T-seal and the speculation is that half T-seal is what floated back in, can you tie this together at all? That blew a crack halfway down the object that came back?

MR. HUBBARD: Well, not anymore than you just did, I think! The float away object on Day Two seems to match its radar peripheral file seems to match two possibilities. One: a piece was one of these panels about this big, about a hundred square inches. Or, you know, a chunk of T-seal, three-quarters or five-eights or something on the T-seal.

I emphasize that we hit, you know, at a chosen angle and force and so forth. If there were 30 percent more, you know, or you had more force to it, could this have cracked across? That's what we're gonna attempt to find out as we do these tests.

And then, seeing what the experiment tells us, then we can go and try to better match that to the Day Two object.

MS. LAURA BROWN: Tracy?

MS. TRACI WATSON: Traci Watson, USA Today.

Dr. Osheroff went into a little bit of detail about what kind of hole you need to get the thermal effects that you saw. Can you elaborate on that a little bit? He said it was bigger than one inch, and then he wasn't quite clear on whether it was smaller than six inches. If you could go into some detail.

MR. HUBBARD: What he's talking about are the thermal dynamic, or so-called aerothermal. It's a combination of aerodynamics and thermal dynamics, calculations being done at different hole sizes to see if the heat then that got in could burn through the wires and create the kind of damage that matches the sensor data. And I don't think that analysis is quite complete yet. I know that they've done some six inch holes.

One of the things that they - and the match with the timeline seems to be pretty good. But, one of the other things that not completed yet is a gap, a slit. It's much harder to do, the analysis is tougher, but they're going to do an analysis to see what would happen if the T-seal were missing. And that was the path for entry. And that result should be available in a - in maybe a week or two.

MS. LAURA BROWN: Earl?

MR. EARL LANE: Earl Lane with Newsday.

Are there any plans to do any arc jet testing on these cracks, either after all the other testing is done, to see what kind of thermal propagation might get through a crack like that?

MR. HUBBARD: There's not a plan at the moment, but it is an idea that has been kicked around. What is in the works is a test of the edge of a piece of RCC, to see if it sharpens up to a knife edge the way some of the debris looks.

There have been a few tests of round holes. It was thought, based on one flight that showed a very small hole in RCC, which appeared to come from micro-meteorite damage. As a result of that, they did an arc jet test with a - started with a small hole and then heated it in an arc jet, and it does expand in a more or less circular fashion, and it forms a very sharp knife edge.

To my knowledge, no one has done a test yet on a crack like we saw, but I can imagine that that would be something people would be interested in. But, that type of test may well go beyond the life of this Board. Those things take time.

MS. LAURA BROWN: Kathy?

MS. KATHY SAWYER: Kathy Sawyer, the Washington Post.

Hi. If you can't make a hole big enough to satisfy yourselves by all of these methods you've just been outlining, is there any likelihood you'll start working things into the scenario such as the windshear during launch, the bolt - the flying bolts from the bolt catcher scenario and so forth? In other words, a combination scenario, or what will you - if not that, what will you do if you cannot figure out how the hole got big enough?

MR. HUBBARD: Well, let's examine just the foam test. It is - you know, I think that we have established the potential failure mechanism. Whether or not we hit it just right is open to speculation. But, getting the crack that we got established that this failure mechanism is at least plausible.

Whether we go from plausible to highly likely, I think, will be determined somewhat by how the rest of these tests turn out, and what kind of analysis you could do to move it around. That's part of the reason they collect this extra data.

For example, if the next test says, well, we get a crack now all the way across, the fact that we have a better analytic tool says, okay, if the actual event were a little bit higher or lower or something, then it would have done this. If we have confidence in that, then we can go from plausible failure mechanism to highly likely. And I think that's the range that we're working in. So that's the answer - part one answer.

Part two answer is, as Admiral Gehman and some of the other board members have said, there are probably some fault tree elements we can never exclude. You can never be absolutely sure that you didn't have a micro-meteorite hit. You can't be absolutely sure that some other debris didn't come off. And those will just remain out there as items that the return to flight people need to consider for future missions.

MS. LAURA BROWN: I'm going to be more liberal about the one question per organization rule here. So, if you guys both want to ask questions, go ahead.

Mark?

MR. MARK CARREAU: I'm Mark Carreau from the Houston Chronicle.

I just kind of want to back up to a basic question on where the testing goes at this point, as you move fiberglass and RCC. What is it that you're attempting to demonstrate at this point? More cracking? Something more dramatic?

MR. HUBBARD: If this was a perfect world, which it isn't, we would be able to establish that we have a very, very probable failure scenario. And that in turn leads to, I mean, knowing what you need to do for a return to flight and knowing, you know, what the origin of this problem was and so forth.

Since it isn't a perfect world, in the sense of being able to absolutely for sure simulate what happened, the best we can do, I think, is to try to bracket the event, demonstrate failure mechanisms, and then understand through analysis what else might have happened.

We've got limited resources in terms of the reinforced carbon panels, and we have, you know, a certain capability to try all these things out. So, that's been out approach, is to try to bracket the event, get some good analysis that compares with the experimental data, and establish where on the line from plausible to highly likely this foam event, as an initiating event, sits.

MS. LAURA BROWN: Alan, do you have a question?

MR. ALAN LEVIN: Alan Levin, USA Today. Sort of follow up on what Matt asked earlier. This at least raises the possibility that you could have relatively minor and difficult to see damage on an RCC panel, but that you'd nevertheless want to know existed while you were in space. What sort of questions does this raise about what kind of tests might be necessary in space and all to determine that the RCC is acceptable?

MR. HUBBARD: The - I think ultimately, it raises the question of, you know, how much vehicle health monitoring do you need to do for a, you know, vehicle that is - still has a lot of developmental characteristics.

This crack was - that we saw and the state that it was produced was visible. It's certainly - the hole crack during the standard investigation, or standard inspection that is done before flight, would've been easily detected, both visually as well as from what they call the tap test. In addition to which they look at all sides of this before - between flights. The interior cracks would've been easily seen. So, in terms - as far as an event, or as far as a defect before a flight, this would've been easily seen.

If it occurred as it did in the accident, on launch, then I think that the unknown yet - we know that there was a breach of sufficient size, which may have been on the order of, you know, inches, that let in enough heat to cause the damage that we saw. So, that size of a thing one would think would be visible from an EVA or some other type of inspection.

The dots - we haven't connected the dots yet from this initial test to the kind of breach that almost surely was present to cause the accident. And that's why I'm saying don't extrapolate too far from the single data point, and come along with us on the, the next few weeks of this R&D activity to see what might have happened.

MS. LAURA BROWN: That doesn't amount to a direct invitation, right?

MR. HUBBARD: Well, I mean, I don't think I've got enough room for all of you, but you can bring your blankets and bedrolls and, you know, whatever else you need.

MS. LAURA BROWN: Bill? Oh, sorry. Wait, wait. I'm sorry. I got the wrong end of the boom mike here. Is there someone over there who wants to ask a question now?

Oh, go ahead.

MS. GWYNETH SHAW: Gwyneth Shaw with the Orlando Sentinel.

I realize this probably sounds like a stupid question, but how did the crack go from being three inches on Friday to being five and a half inches now? Did you take two measurements? Is it the mix up with the centimeters? Is there a possibility that it grew after the first time you looked at it? I mean, I'm just curious.

MR. HUBBARD: No, first of all, there are no stupid questions.

What you got on Friday was my quick, you know, minutes long, you know, 30 minutes or so, investigation of a very complex test reported to you as rapidly as possible so that you'd get some feel for what happened.

I hope I was clear at the time that this was a very preliminary result, based on a quick inspection and promised what we're doing today, which was to come back after we had had time to look at it much more carefully and go and look at all the sensor data and take actual measurements, as opposed to, you know, holding a little pocket ruler up to what you can see inside this thing. I mean, they actually took this whole thing apart, the T-seal, the panel and everything, took it into the shop and were able to make very careful measurements.

So, you know, it's just the difference between, you know, going into your, you know, kitchen quickly and looking for your car keys or something and - or you know, trying to find the right utensil and you do what you can in a few minutes versus taking the time to go and do it carefully. And that's the only difference.

MS. LAURA BROWN: Okay. Bill?

UNIDENTIFIED MAN: So I just want to be clear. We have two new cracks since last Friday.

MR. HUBBARD: We have - what I reported Friday during the press conference was the crack that is now, by careful measurement, five and a half inches long. It appears as a three-quarter inch crack on the surface.

Later that day we found the crack on the T-seal and we put that in the press release, and that was about two and a half inches long in this area here. And then, after even more careful investigation - and I also at the time reported that this panel had been shifted a little bit, gave some approximate numbers, and we found that there was a chip on the carrier panel.

What we've added since then is the crack up here near the flange area, where the bolts attach this to the structure. So I believe, if I've counted that up right, we only have actually added one new piece of damage.

UNIDENTIFIED MAN: So given then the scope of the damage, can you characterize just generally how the panel sustained or didn't sustain the damage from the impact?

MR. HUBBARD: I'm sorry. I didn't follow you on that one.

UNIDENTIFIED MAN: Did it hold up better than expected or worse than expected?

MR. HUBBARD: The pretest predictions were all over the map, from a simple crack, which is what we observed, to breaking into pieces. And it's because nobody has ever done this kind of testing before that we are in this, you know, research development type of activity.

The - what we saw is more at the lower end of the predictions. But, as we peel the onion and look deeper into it, we see that some things match up very well and other things are somewhat of a mystery. Like, how could we have a force out here three times the breaking strength and nothing happens? So, that's why we're conducting more data.

So, in general, I think we got a pretty good match with one set of models, but there are still some question marks.

UNIDENTIFIED MAN: I'm gonna ask you a question unrelated to the immediate subject at hand, which is - that I assume you know the answer to. The explosive bolts and the bolt catchers we were talking about earlier, can you just explain in a very simply way how that mechanism works? I mean, if the bolts get split, is there a thing that blocks it? If it gets moved out of the way mechanically?

MR. HUBBARD: What - you have a bolt holding the solid rocket booster onto the external tank. And when the bolt, explosive bolt explodes, it separates, and one piece goes with the external tank, one piece goes with the solid rocket booster. But, it's just a mechanism of cutting it, you know, so that - .

UNIDENTIFIED MAN: - So the bolt actually gets cut.

MR. HUBBARD: Yeah. Just cut it. These things are often called “bolt cutters” in the technical parlance.

MS. LAURA BROWN: Ricardo?

MR. RICARDO ALONSO-ZALDIVAR: Yeah, Ricardo Alonso-Zaldivar with The L.A. Times.

Okay, this is kind of like an arcane question, but I was just wondering. That tiny little three-quarter inch crack on the surface, right? I just wondered, did you calculate what the space is, like how many square inches that would be, the opening, to give us an idea? You see what I'm saying? The gap?

MS. LAURA BROWN: You mean how wide is the crack?

MR. ALONSO-ZALDIVAR: Yeah.

MR. HUBBARD: Oh, no. Haven't calculated that yet. And if you remember in the video there it actually, depending on how this thing flexes, it opens and closes, you know, the crack itself is the sort of thing you'd slip a piece of paper into and it's not very wide. But I don't have a number for you, off the top of my head.

MR. ALONSO-ZALDIVAR: (Inaudible).

MR. HUBBARD: Ah, interesting point. The - when it's closed, a quick look didn't show any light shining through it. Somewhere - and we didn't - you didn't get to see the full video - somewhere in that impact, something opened up enough to let light into the channel here, the inside of the leading edge. Whether it was the crack or not, I don't know. It could've been as simple as the T-seal moving aside and then going back together again.

But, the point is, the point is not whether you can see light through it, and it's not how - it's not how wide it is. It's the fact that you have penetrated the thermal protection barrier. This crack goes all the way through the rib. The thermal protection of this material is provided by the silicon carbide, which is a part of this material. This material is made in a very complex process and, essentially what happens is, they really convert the carbon on the outside into silicon carbide, which is the actual protecting material. And we have a crack that wraps all the way around, that goes - that penetrates the thermal protective material.

MS. LAURA BROWN: Okay, do we have other questions?

Bill?

UNIDENTIFIED MAN: Just a scheduling question. Scott, you said you're leaning toward maybe using a real panel eight. The schedule you gave for shooting this panel eight through ten, would that support - if you're using a real one or if you decide to use the real one, would that test slip into July?

MR. HUBBARD: The three or four options that we're looking at all support a panel eight through ten test by the end of June.

UNIDENTIFIED MAN: Is that something we'd be able - be able to come to if we wanted to?

MS. LAURA BROWN: I guess we'd have to work that out. I mean, it depends on a lot of different factors at that point. But I think we would - .

MR. HUBBARD: - We'd lean in the direction of doing what we did before for the real RCC test. I don't think you're - you know, you need to spend your time on the fiberglass run-ups and all that. But, for the real test, you know - .

UNIDENTIFIED MAN: (Inaudible).

MR. HUBBARD: Yeah. We'd try to do it if we could work it out, like we did before.

MS. LAURA BROWN: Okay, other questions? Okay, we've exhausted you! Finally!

So, all right. Thanks everybody for your patience and - .

UNIDENTIFIED MAN: (Inaudible).

MS. LAURA BROWN: That's the last hearing that we have - currently have scheduled.

UNIDENTIFIED MAN: (Inaudible).

MS. LAURA BROWN: We are considering a press briefing in two weeks. We have a schedule with the Board right now where, since several members of the Board are on the West Coast, we're getting the full Board together for several days together, every other week. And so we would probably not do a briefing next week. But there's a possibility we would do one two weeks from now.

UNIDENTIFIED MAN: Here?

MS. LAURA BROWN: I don't know if it's gonna be here or not. We have - if the room is available, we may be doing it in one of the - in this area. We have several other facility options, too. But, it would probably be in this part of town.

UNIDENTIFIED MAN: (Inaudible).

MS. LAURA BROWN: Yeah, exactly.

Thank you.


END

// end //

More status reports and news releases or top stories.

Please follow SpaceRef on Twitter and Like us on Facebook.

SpaceRef Newsletter