Failure In An Expected Pattern

by Don Friedman on September 20, 2017

The animated picture above is from The General, a 1926 Buster Keaton movie. (Click on the picture to enlarge it.) In the era before computer graphics, and when the use of scale models for special effects led to disappointing results, Keaton got a realistic train crash in the obvious manner: he crashed a train and bridge. This is one of the reasons that the movie was, for its time, considered incredibly expensive.

If you watch the clip frame by frame, or simply watch it a few dozen times, you’ll see the way the bridge fails: the left end lifts up off its vertical supports and then breaks, triggering the failure of the main span. This raises the question of why? Why does the bridge fail in that particular manner? Common sense – which is, of course, so often wrong – says that if the main span is overloaded that should fail first.

The bridge, probably by design but maybe by accident, is continuous across the left support, meaning that bending moments and shears are transmitted across that support. One of the interesting results in analysis is that the “negative* bending” moments at supports tend to be higher than the positive bending moments at midspan. So, if the bridge was of equal strength and stiffness along its length, as it probably was**, then it would deflect fastest at the continuous support and break there first, which is exactly the behavior we see. Using the same logic, the righthand support is probably not continuous, as it simply bends down as the main span sags and then breaks after the collapse has already begun.


* A fellow student during my second year of college introduced me to a mnemonic to help remember which moments are positive and which are negative. Positive moments cause a beam to bend downward in the middle and upward at its ends, so that it looks like it’s smiling; negative moments cause a beam to bend downward one either side of an unmoving support, so that it looks like it’s frowning. This is an oversimplification of the difference between positive and negative moments, but it’s also unforgettable.

** This was an old and primitive bridge for a branch railroad. It’s easier to design and build a girder bridge that is the same strength for its whole length than it is to design one where the strength reflects the requirements at any given point.

The Turning Point In Frame Technology

July 5, 2017

Ben Evans has a very nice piece on the evolution of technology here: Not Even Wrong. It’s not a completely new idea, but he states it well: there is evolution of a given technology within a type, and there’s development of new types that require new ways of design and analysis. These two kinds of […]

Read the full article →

Jurassic Silliness

May 20, 2017

I’m not sure Jurassic World needed structural analysis, but it’s got it now.

Read the full article →

Honest Critique

February 11, 2017

Not a McMansion: Architects are first exposed to critique in school, where students discuss the various failings of each other’s work. I witnessed this as an outsider and was astonished at the extent of criticism. Nothing prepared me for McMansion Hell, Kate Wagner‘s website that eviscerates overlarge suburban houses on a weekly basis.

Read the full article →

Arching Action, Too

January 31, 2017

One of the odd administrative aspects of running a blog is that I get statistics on how people arrive here. A blog post from last spring, “What Is Arching Action?” is now our third-most popular post, with a few people arriving here almost every day after searching for those words. (As of today, that blog […]

Read the full article →

This Is Not An Architectural Critique Of The New PATH Station, Part 2

January 4, 2017

Trippy: (Click on any image to expand.) In part 1 (last March!) I mentioned that I was interested by what I could see of the roof of the PATH station itself, which is due west of the Oculus. The Oculus has received most of the press, is visible aboveground, and is (by virtue of the […]

Read the full article →

A Symptom Only

November 15, 2016

Another picture from Gabi’s visit to a church. Those cracks look awful, but they are the equivalent of a fever: they are the response to a problem rather than the problem itself. In a true gothic-built church like this* the vertical buttresses, flying buttresses, piers, and arches make up a stone skeleton, and the walls between […]

Read the full article →

The Inflexible, Bent

November 14, 2016

We informally divide the world into flexible objects and inflexible ones. That’s a gross simplification based on a bad idea. Everything is flexible; the only difference is degree. The example that I have used roughly one per year since the late 1990s, while teaching, is that the Great Pyramid of Giza bends when the wind […]

Read the full article →

Incompatibility: Cavity Walls and Veneer

October 17, 2016

How McKim, Mead and White handled a solid wall, with brick shown as diagonal hatching and stone as stipple hatching: I previously talked about the problem of providing support in cavity walls and how that is inherently different from the way in which masonry walls evolved. That’s actually the lead-in to a bigger problem, which […]

Read the full article →

Incompatibility: Loading

October 12, 2016

New York tenements with changes of use at the first floor. I’ve emphasized reuse a number of times. The phrase that you usually see is “adaptive reuse,” implying a change in occupancy, because ordinary reuse of buildings happens all the time and is rather boring. When a house is sold from one owner to another […]

Read the full article →