Disclaimer: I’m not an engineer. What follows is my own opinion, derived from newspaper articles and casual rubbernecking stemming from a fascination with large machines and structures.
Those of you reading this from outside of the San Francisco Bay Area may not be aware that one of the major bridges that crosses the Bay is being replaced. Construction has been going on since 2002, and the new bridge is expected to open this September. For background, see the obligatory optimistic official info site. Keep in mind that this bridge is being built as a result of seismic safety concerns stemming from the 1989 Loma Prieta earthquake, but construction was delayed by multiple arguments over the design: whether to build a cheap, serviceable structure, or a regional showpiece.
As with any large construction project, this one has had its share of problems. I doubt anybody sane would have expected the project to be trouble-free, and it seems to me like the problems so far have been within the range of trouble that should be expected for a project this size.
The latest problem, though, has set off my internal alarms. Per a series of stories in the SF Chronicle (see, for example this one and this one), a key part of the seismic safety system are a set of devices called “shear keys”. These devices are connected to the bridge structure with a set of large bolts, 288 of them. The bolts were installed before the the roadway was constructed; now that the roadway is in place, the bolts are being tightened. Of the 96 that have been tighted so far, 30 have failed, fracturing when put under tension.
Given that it’s still early in the investigation of the problem, there isn’t a whole lot of information available. We’re still in the stage where everyone says “We’re looking into it,” and haven’t reached the point where everyone starts saying “It was the other guy’s fault.” That being the case, I’ll skip over the fact that (at least according to the Chron’s articles) nobody seems to quite know who actually manufactured the bolts. I’ll also skip over the question of how the bolts were tested (the articles say they were, but don’t give any details).
What I do want to comment on is the aspect of the situation that makes absolutely no sense to me: the bolts range from 9 to 24 feet long, yet there is only 5 feet of clearance beneath the roadway. That means the bolts can’t be replaced. Again, I’m not an engineer – or an architect for that matter – but this seems like a bad design. Even if the bolts were not defective, wouldn’t they fail eventually? I’m sure there’s plenty of safety margin built in, so the bridge wouldn’t be in any particular danger if a few of the bolts failed, but wouldn’t there be a need to replace them at some point? I haven’t seen anything that suggests the new bridge was designed with a limited lifespan like a phone with a non-replaceable battery. Battery dies, buy a new phone. Earthquake stresses the shear key bolts, buy a new bridge. Isn’t the whole reason we’re buying this bridge because the old one needed replacement after an earthquake?
Maybe there’s a valid engineering reason why the bridge was designed this way, but right now it doesn’t seem reasonable. I can’t help but think that when “regional showpiece” won over “cheap and serviceable”, inadequate attention was paid to the reason why a new bridge was needed in the first place.