August 5, 2007

Inspection Techniques Under The Microscope

Bridge inspections have come under scrutiny after the collapse of the I-35W bridge over Mississippi, with Minnesotans wondering how a bridge that passed an inspection in May could collapse less than three months later. The technology involved in bridge inspections may surprise some, as they still rely on hammers, steel chains, and eyeballs for most of the analysis. Also, the design itself may have been the major contributing factor, according to a former New York City bridge safety engineer:

As canoeists paddled below, state bridge inspector Eric Evens stood in a cherry picker next to the two-lane steel bridge over the St. Croix River near Scandia on Friday, eyeballing rust, cracks, bolts and rivets.

Computer-aided design and other innovations have changed the way structures are built, but bridge inspections haven't changed much over the years. In this high-tech era, the trained eye and the rap of a hammer to listen for the sound of bad metal are still considered among the best ways to examine bridges.

Many inspectors insist that the great majority of bridges are safe and that their work is thorough. But others say the failure to predict last week's catastrophe in Minneapolis points to flaws in the system and the need for better technology to detect problems.

It turns out that bridge inspections are more art than science. The inspectors listen to the sounds that bridges return from hammer taps and chains dragged across girders more than using high-tech measurements of density and corrosion. Most of the normal routines rely on visual inspections -- eyeballing for corrosion, cracks, and fatigue.

How effective is that? In 2001, the Federal Highway Administration decided to test inspectors on a subtle but significant flaw in a bridge. Only 4% found it.

Technology may be a limited solution, however. The Admiral Emeritus spent years in the space program in quality engineering. He held the highest level of expertise in non-destructive testing, which was used on the Shuttle program to test support structures for engines and the like. That kind of testing was difficult enough in a hangar designed for access to the spacecraft's interior, but would be much more difficult to apply to bridge inspections in the field, especially when inspecting a bridge like the St. Anthony Bridge that collapsed last week.

The Star Trribune points out another issue that led to the collapse. The bridge's design and construction in the 1960s is a potentially critical factor, not for its age but for the philosophy of design at the time. Samuel Schwartz, who closed his share of bridges in New York City, explains that bridges built before and after that period have redundancies designed into them. Not so with the post-war period going into the 1960s, however, when American builders and government officials felt that bridges had been overdesigned and excesively costly. After a number of collapses in the 1970s, the designs went back to redundancies -- but apparently no one thought to add to the existing bridges, even those high-traffic structures like the I-35W bridge.

Another change from the time of the design also may have been a factor. As late as the 1960s, rail traffic handled much of the transportation of goods to market, and trucks were used for mostly local distribution. That has changed in the 40 years since, and heavy vehicles make up a larger percentage of highway traffic. The assumptions for use made in the 1960s stopped baing valid two decades or more ago. The heavier traffic going across the bridge would accelerate fatigue and bring failure even closer.

With all of that, though, the question still remains why the inspections did not catch whatever caused the failure. That question cannot be answered until we know what specifically did cause the collapse, and that may take months to learn. Until then, we have to consider what to do with bridges of similar design and age, and figure out how to inspect them properly.


TrackBack URL for this entry:

Comments (4)

Posted by vet66 | August 5, 2007 10:22 AM

Let's include railroad bridges in this equation. Unit trains of coal and grain generally weigh 10-16 thousand tons, not including locomotives, running across bridges that in most cases are approaching the century mark. Trucks may be hauling increased amounts of traffic on highways but so are railroads as intermodal freight has taken off.

Think of that the next time you drive under a railroad trestle with a train moving across it at speed. You can usually reroute barge and truck traffic around problematic bridges. It is an entirely different matter when it comes to rerouting trains on a rail system already maxed out.

Posted by RBMN | August 5, 2007 10:26 AM

It may turn out that the bridge’s problem was just poor quality control during the manufacture of some of the steel, some of the joints, or even part of the foundation--parts that random inspection happened to skip. The inspection depends on the assumption that the starting condition of the bridge was sound. That may not be the case. Rivers also have lots of power, and do things (over time) that inspectors can’t see or even anticipate. I think the thing that today's investigators won't find is bad faith. Nobody in charge ever thought they were just throwing the dice to allow this bridge to be used. That's not Minnesota.

Posted by athingortwo | August 5, 2007 10:58 AM

Captain - Bridge inspections cannot ever be expected to prevent all possible structural failures. Not in the real world. What bridge inspection programs CAN do is reduce the incidence of major unforced structural failures. The effectiveness of the bridge inspections system that we have now can indeed be demonstrated by the fact that the Twin Cities I35 bridge is the first major highway bridge structural failure in the USA not caused by a collision with a ship or barge, or major earthquake in the US since the Schoharie Creek bridge failue in NY back in 1987 - 20 years ago.

In the case of the I35 freeway bridge, it made for a few days of spectacular video at five, and much wailing and gnashing of teeth by all manner of politicians and special interest groups with an ax to grind, but in the overall pantheon of famous structural failures, the loss of life (max of 13, if all the missing persons turn up dead in the river) in the Twin Cities last week is unremarkable in comparison with the daily carnage on the highways we all know about personally.

Add up all the highway bridge failures in the USA in the last 40 years (even including those caused by collisions with vessels and earthquakes) and the number comes to 226 fatalities .. or fewer than 6 fatalities per year on average.

In 2003, US fatalities due to vehicle collisions with animals (mostly deer) was 201, and the annual average is over 100 human fatalities and 7,000 injuries (not to mention 350,000 deer fatalities).

Which means that the average driver is nearly 20 times more likely to be killed by a deer on our highways than by a falling bridge.

Yet, I don't recall any big political campaigns for increased spending on deer barriers, or moose control, on our highways, do you?

A driver or passenger is also 2,886 times more likely to be killed by a drunk driver than by a falling bridge. Yes, we as a society are actively trying to reduce that risk, but most people still travel our highways without giving much thought to the risk of drunk drivers - even when faced with daily if not weekly direct personal experience of drivers who appear to be under the influence of alcohol. And we don't talk much about that fact, nor do we run continuous breathless video reports on the 24-hour cable news channels about it either.

Let's get real, folks.

Posted by cirby | August 5, 2007 3:43 PM

"Yet, I don't recall any big political campaigns for increased spending on deer barriers, or moose control, on our highways, do you?"

Isn't that what the NRA is for?

Post a comment