The Phoenix > NEWS > Spectacular failures

Spectacular failures


An MIT instructor on the Big Dig disaster

By: ANA RIVAS

7/25/2006 3:06:12 PM


The Big Dig
Engineers have taken us out of the caves and up to the skies, linked us with far away shores, and connected us to distant lands. And along the way they’ve also made some pretty huge mistakes. Stephen G. Banzaert, an instructor at MIT’s Edgerton Center, teaches a course called “Spectacular Failures in Engineering and Other Experiments.” ThePhoenix.com checked in with Banzaert to see if the recent Big Dig disaster qualifies as a spectacular failure, and to find out his speculations on what, exactly, went wrong. What follows is an edited transcript of the e-mail conversation.
Would you consider including the Big Dig on your syllabus this fall?
Given how long I've now spent on the phone talking about the Big Dig, I think it's almost a certainty that it will be on the syllabus this Fall, although it may yet be too soon to draw any real conclusions from it.

What are the "Spectacular Failures" that you teach in your class?
We range from the Boston molasses flood, to an explosion of a gun on the USS Princeton, to both shuttle disasters, the THERAC-25, and the Bhopal chemical disaster. We focus primarily on disasters that present lessons to engineers and organizations on how to do better in the future.

What took you to teach your students the bad example?
My class is geared towards freshmen as an introductory course in the history and causes of engineering failures. Last year I had 20 students. It's a new version of a class I took as a freshman at MIT (class of '98), which had left the catalog a few years before I returned as an instructor. I started teaching it because I think engineers in general benefit from learning this history, and perhaps more importantly from learning how to recognize the environments that allowed some of the "classic" failures to happen. As I mention in the description of my class, it's a humbling proposition to think that most failures are caused by motivated, qualified, highly trained people who are doing their best and think they are succeeding. It's hard to see your own blind spots — hence, for example, the need for external review.

Can you shed some light on what went wrong with the Big Dig tunnel’s ceiling?
I have no idea what went wrong. Anyone who has direct engineering knowledge of those ceilings is, for now, not talking. Until we really understand the mechanism, there's no telling what happened.

The first bit of information that came from Big Dig authorities was that a hanger bolt failed and that caused the ceiling to fall. The question of the (early) hour was: Did the failure of one hanger bolt actually cause this collapse, or was there a larger network of failures (such as many weakened hangers, with one actually failing and precipitating the collapse)?

Typically, designs of this nature come with substantial redundancy such that the failure of a single member can't cause a collapse like what we saw. (For example, your car doesn't lose a wheel if a single nut comes unscrewed. You would either need several nuts coming off, or all of the nuts to be weaker than specified, in order to be able to rip a wheel off.) There's no telling what the actual scenario here is until some skilled structural engineers spend time poring over the blueprints.

Now the Governor and his engineering team are saying that none of the 1,146 epoxy bolt hangers can be trusted… maybe all of the nuts were weak after all. Do you know what this "epoxy" glue is, where else it's been used? Do the panels HAVE to be so heavy?
Unfortunately, I have no idea where else this technique is used. What I can say about other ways of building and the weight of the panels is: In any design process, you would consider all the options that met your design criteria and winnow them down until you were satisfied that you were looking at the best option. So at some point, I have to assume that all the alternatives that are being tossed around (use lighter panels, hang them some other way, etc.) were considered and rejected.

What's somewhat critical in understanding the process is for us to eventually find out what was rejected, why, and furthermore what were the design criteria in the first place? (As a simple design example, the decision to use 3-ton panels obviously carries with it a new issue — namely, that the hanging system needs to be basically infallible, since the consequences of a failure are tragic rather than simply inconvenient. So the question is, what do you gain by having panels that heavy, so that you're now willing to take on that new requirement? Presumably there's a good reason, but I don't know enough about the inner workings of that tunnel design to start speculating on what it might be.)

Can you speculate if it was a matter of bad math, too-fast construction, or too cheap materials?
Again, no idea. Could be any or all, or some combination, or something else entirely that nobody's thought of yet. Some failures can be traced back to a single batch of bad steel; others are caused by long chains of failures, the prevention of any one of which could have stopped the whole event

Would you drive through the broken tunnel?
Of course I'd drive in the tunnel. Just like I still drive a car, even though they've had failures in the past that have now been fixed. I have to trust that those responsible will learn from whatever happened and deliver a re-opened tunnel that's safer than it was before.