Minneapolis I-35W Bridge Collapse: A Truss Failure?

Check in with our I-35W bridge collapse coverage page for the latest news, videos and photos covering the failure.

I am not a structural engineer, but I did some groundwork on bridge history and problems with the I-35W bridge. I don’t want to rehash what is probably common knowledge, but a 2001 University of Minnesota report found distorted girders and signs of fatigue, and that the bridge could collapse if part of a truss gave way.

Very telling and alarming to me was a real-time image of the collapse of the span over the river. I have not seen an image of the intact bridge, nor do I know the actual bridge measurements I have used to describe my observations, but here goes. In the image I saw, during the playback interval, it appeared to me that the collapse initiated at two places — each place across the bridge deck about ¼ of the length of the river span toward the span center. During the video playback interval, the span over the river appeared to be already falling and appeared to be below the level of the bridge deck at the piers. 

My opinion is that the trusses had failed, or were failing at the time the video was recorded, and allowed the central half span to sink towards the river.   The bridge deck of that span looked level and flat, falling straight down.  Furthermore, the images I saw showed the ¼-length spans at each end sagging downward while still attached at the bridge deck at its associated pier. Once the central part of the span fell far enough, the short, ¼-length spans pulled from each pier and also fell.

What initiated the truss failure? I don’t have the benefit of a metallurgical analysis, but my opinion is that corrosion could have played a part. This bridge is 40 years old and I think a design-related stress failure would have shown up sooner. This bridge carries over 140,000 vehicles per day, with many trucks in the mix of traffic. Corrosion could have weakened truss joints and the constant pounding of heavy traffic may have triggered the failure I described above. I’m sure that as the investigation goes forward, we will all gain an understanding of the specific cause of this failure.

My last thought, and it may not be original, is that when bridges, or other life-critical structures are built, such structures should have strain gauges applied at critical joints and spans at the time they are built. These strain gauges, a permanent fixture, could be wired to a central location and be periodically scanned for a readout of bridge health, if you will. These gauges could also give an opportunity for a real-time readout as vehicles cross the bridge, or during seismic or other stress.