Replies

I got an e-mail asking some additional questions and sent the following reply. If anyone else has the same questions, here it is:

Y”ou may be interested in the article I wrote in FreeRepublic.com after I completed my last refresher course in bridge inspection. It covers some of your questions. It is at:

http://www.freerepublic.com/focus/f-bloggers/1598099/posts

Since it lasted as long as it did, I do not believe that there was any “bad” steel or “bad” concrete used. I think the primary reason that it did not last longer was the design. This is not to say that the original designers were incompetent or evil. They were going into unknown territory at the time.

There were a huge number of new bridges constructed as a part of the Interstate system that got started in the late 1950’s, through the 1960’s, and into the 1970’s. This was designed in the early-1960’s, started construction in the mid-1960’s and was finished in 1967. Because of the large number of bridges needed, they had to be built differently from the past. Less cost, less material, and more speed. This led to the adoption of many things that should have been tested out more thoroughly before use. I touched on several examples, but here is more detail.

The problem of fatigue was unknown in bridges before this. Fatigue was known in WWII, but was thought to be limited to airplanes. Unlike airplanes, bridges were big, heavy, used low strength steel with lots of ductility and plenty of excess material for corrosion. Also, the earliest bridges (stone arch) were totally in compression where you cannot have fatigue. However, steel allowed the use of tension. By using both tension and compression, bridges became lighter. However, bridge designers did not consider fatigue problems when designing welding details. They just used whatever was the quickest and easiest connection details.

Some of those welding details have directly caused fatigue cracks in this bridge. Things like tack welds, skip welds, plug welds, back-up bars that are not removed after final welding, are now known to be BAD, BAD, BAD for welded structures. However, usable design standards for fatigue were not widely known in the industry until the mid to late 1980’s, long after this bridge was designed. Other fatigue problems were with design. Diaphragms without without stiffeners on the other side of the web or stiffeners that just ended in the middle of nowhere are bad fatigue design details.

The mere fact that welded box beams were used is also a fatigue problem. It looks like large parts of this bridge were welded somewhere else and trucked to the jobsite, where they were joined by bolted connections. This speeded up construction and was a lot cheaper, but a welded structure, once a fatigue crack starts, will crack completely through. In earlier times, when a riveted box was used, a crack would only go through ONE of the four parts of the box and would stop at the riveted joint.

The next thing was the move to higher strength steel. That means that the material used is thinner, has less ductility (which means less fatigue resistance), and corrodes just as easily. If you lose 1/4” of a 1” thick member it has lost at least 25% of its original strength. However, if a higher strength material only needs a 1/2” thick member to begin with, the same amount of rust will reduce the strength at least 50%. However, high-strength steel was looked upon as a godsend for reducing cost at the time. There is now something called “corrosion resistant steel” (A588 and the like), but it is not good for places that move (such as bolted joints and expansion joints) where a lot of the problems with bridges are.

I also mentioned that the old computers (IBM 1620 was the first one I used) first came into being about that time, along with calculators, which allowed engineers to convince themselves that they could use thinner material. Again, thinner material is also more susceptible to corrosion and fatigue.

Keep in mind that once fatigue cracks are seen at least 90% of the life of the structure (at that point) is gone. A structure with extensive fatigue cannot be repaired to be “good as new” no matter what is done to it. At the very best, you can extend the life SLIGHTLY. They tried to extend the life of this one too much.

In other words, it was the perfect storm – welding, higher strength steel, lack of knowledge about fatigue, and engineers with computers. I don’t believe that anything was done deliberately that caused the collapse other than the refusal by someone to believe that it was as bad as the inspectors said and the subsequent refusal to speed up its replacement.”

Thanks for you analysis. It appears to fit the pattern discussed in a book on engineering disasters I read recently.

The book asserted that bridge building follows a pattern — after a status of using known materials and methods to yield safe bridges, new materials and processes become available. The first builders to adopt these new materials and processes grossly overdesign the first bridges, and they are safe. Then as time goes on, they gradually remove the safety factors in the interest of lower cost and/or building time (e.g., using calculators to convice themselves that the thinner high strength steel will carry the load). Then a disaster occurs, and engineers figure out the problems, and change something to deal with that problem so that the bridges are safe again. Then .. new materials and processes pop up, ... and the cycle repeats.

It would be facinating if it didn’t result in death.