I keep talking about how engineering moves forward by looking at past mistakes. Sometimes the mistakes are caused by ignorance, in which case it’s often possible to find the event that triggered research to reduce the level of ignorance. For engineers in the US, the St. Louis tornado of May 1896 was an event that led to much more serious research about wind pressure. (Unfortunately, this kind of lesson doesn’t always cross national borders. Many of the lessons of about wind pressure could have been learned from the Tay bridge collapse in 1879, but Americans didn’t change practice significantly until there was an event here.)
The tornado killed over 300 people, destroyed more than 300 buildings, and damaged at least another 7000 buildings. Julius Baier, a St. Louis engineer, took a good look at the damage caused to different types of buildings and what it said about the pressure exerted by the wind. Here is an example of a railroad building that had a steel truss roof supported by masonry walls:
Baier’s article, “Wind Pressures in the St. Louis Tornado, With Special Reference to the Necessity of Wind Bracing for High Buildings,” published in the Transactions of the ASCE, is far more interesting for his analysis than for his descriptions of damage. He spends a fair amount of time looking at the formulas used to convert wind speed (reasonably well measurable in the 1890s) to wind pressure; in his conclusion he looks at a number of specific building types and discusses how they performed.
All of the discussion is worth reading, but for my work, I found the most interesting to be his conclusions about tall buildings. As the subtitle of his piece indicates, this was a topic of interest for him and for the engineering community during the first big wave of skyscraper construction. In 1897, there was still some debate as to the best way to construct a tall building: with a cage frame, where the metal frame was braced by the masonry of self-supporting exterior walls, or with a skeleton frame, where the frame was braced by diagonals or rigid connections within the metal. Baier concluded that skeleton-frame buildings are superior to cage-frame buildings because the loss of exterior-wall masonry won’t destabilize the skeleton. This may be the first analysis-based statement by an engineer that there was a serious structural flaw in the cage-frame concept, as opposed to earlier analyses that focussed on failures of cage frames as expressions of construction or design errors. (I’ve put his report in the context of other forensic reports here.)
Finally, the picture at the top of the page is from Baier’s article but it’s not from St. Louis. It’s a waterspout – a tornado over open water – near Cape Cod. Tornados are not really the best way to look at wind pressure for structural design, but you work with what you have, and what Baier had was the St. Louis wind records and his observations of damage.