Part of the Construction History Congress last week was a day of tours, and I went on the iron- and steel-truss bridge tour. I’m going to show my pictures every two or three days, spread over the next couple of weeks. First up: the rail bridge at Eglisau:
Built 1895-1897 by the Swiss Northeastern Railway, designed by Robert Moser, and a combination of ashlar masonry in the approaches and steel truss for the main span. It;s 50 meters above the river, which is nice and dramatic. It’s fun to take pictures of the approaches
but we were there to look at the truss and that’s the topic I have semi-deep thoughts about. The truss is a single-span Pratt deck* truss, with diagonals that typically cross two panels rather than the more-ordinary one panel. This is a double-intersection Pratt truss. That style is more difficult to build than a single-intersection truss and is not statically determinate, although there are ways to simplify the analysis if it’s the 1890s and you don’t have a calculator. The obvious question is why anyone would use this form rather than a single-intersection truss and the answer is that design involves making tradeoffs.
The starting point is the length of the main span, 90 meters. It was set by the width of the Rhine River, as the last two approach piers are more or less at the opposing banks. The span and the design load (the weight of a train) gives a first-pass guess at the depth: the maximum moment at midspan has to be carried by the top and bottom chords, so limiting the stress in those members and the deflection overall will suggest a depth. Once you’ve got that, you need to set the panel width. Ideally, you’d like the panel width to be close to the truss depth, as 45-degree diagonals are more efficient than other angles, but that causes a problem if the truss is deep. The beams supporting the rail-bed are not true truss members, which only carry tension or compression. Rather, they’re beams in bending and their design is extremely sensitive to the span between supports, which is the width of the panel.
So to keep the rail-supporting beams a reasonable size, the panel points get pushed closed together. And then to keep the web diagonals at a reasonable angle, the design gets pushed towards double-intersection.
* This isn’t a true deck truss, where the rail-bed is supported by the top chords. Instead, the rail-bed is downset slightly and supported by a set of beams fastened to the web verticals. (We were told this configuration was used because it was safer: a derailed train would not immediately fall into the river.) That said, the issue described is the same for this variation as for a true deck truss.
You must be logged in to post a comment.