I seem to keep coming back to train station roofs as I write. That’s happening partly because I travel by train a good amount so I’m looking at those roofs, and partly because they tend to have exposed structure designed by engineers with little or no separate architectural design, which may not make them beautiful but it does make them good fodder for an engineering blog.
That’s a piece of platform roof at the Hartford station. It was built in 1889 and repaired in 1914 after a fire; a photo from before the fire shows platform roofs similar to these so the steel structure seen here is either original and 128 years old, or it’s a replacement and only 103 years old.
The main structure, as seen above, consists of a series of gable trusses spanning the short direction over the platform. Every third truss is directly fastened to the pairs of columns; the other trusses are simpler and are supported by parallel-chord trusses running the length of the column and supported by the same columns. The wood roof deck simply spans from gable truss to gable truss. The knee braces create moment connections between the columns and both the gable and the straight trusses, allowing the roof to resist wind load. All in all, it’s a logical design that seems to have performed well for over a century.
The big question: why such a complicated form? Those trusses require a lot of pieces of steel and a lot of riveted connections. Engineers today would never use such a design unless forced into it by an architect or owner who wanted an oldey-timey look. The reason is that the cost in steel structure today is in the labor for fabrication and erection, and having all of these little pieces and connections means it’s a lot of work and therefore expensive.
Things change. In the 1880s, and arguably still in the 1910s, the biggest cost in steelwork was the steel. Engineering design had reducing the poundage of steel as a main priority to keep costs down. These trusses weigh less than a similar structure constructed using, say, I-beams would weigh. As a matter of fact, I’m hard-pressed to think of a form of steel structure that would use less weight of steel than this. Those trusses are very efficient on that metric.
The old design is efficient on steel weight; a newer design would be efficient on connection labor. There are other criteria you could use to measure a steel design: for example, exposed area of metal (which matters for weathering speed and amount of painting required), or piece weight (which affects the equipment needed for erection). All of these count as efficiencies and all contradict one another, which makes it impossible to maximize more than one efficiency at a time. Telling a structural designer to “be efficient” is meaningless. You’re asking for some criterion to be maximized or minimized, but which one? Make your pick and then you get to live with the results.
You must be logged in to post a comment.