Windswept_ The Story of Wind and Weather - Marq de Villiers [55]
Among the most important case-studies of local winds are those that involve long-span bridges. These are very complicated structures, susceptible to wind in many ways. They are liable to swaying, and to oscillations; the cables are liable to quiver dangerously, like massive violin strings, in high winds; and all components undergo stress and therefore fatigue. Oscillation is the enemy of bridges, and engineers must install what they call vortex-dampers to head it off. Out-of-control oscillations have destroyed a number of such structures over the years, including the 1836 collapse of the Brighton Pier in England, the 1879 collapse of the Tay Bridge in Scotland, the 1940 collapse of Tacoma Narrows Bridge in Seattle, and the 1986 Amarube Tekkyo rail bridge in Japan. Perhaps the most notorious of these was the Tacoma Narrows Bridge, which was captured on film—and which can be viewed on any of dozens of Web sites run by disaster junkies. The most interesting result of the collapse was to cause a new note of sobriety to enter the world of bridge design. Since the nineteenth century, bridge engineers had been besotted with new techniques and new designs, and suspension bridge designers competed with each other to achieve a maximum of structural grace and slenderness; artistic merit more or less overshadowed conservative engineering—a fact of life now endlessly drilled into the skulls of first-year engineering students. With its very shallow trusses and slender towers, the Tacoma Narrows Bridge was the high point in bridge artistry Alas, it liked the winds far too much. It was shaped not unlike an aircraft wing—except that bridges are not supposed to generate lift. Only a few weeks after opening, the bridge had already developed the nickname Galloping Gertie for its tendency to heave in even moderate winds. A gusty windstorm in November 1940 was enough to do it in altogether. For an alarming few minutes, it twisted violently in the wind—at maximum twist one sidewalk was twenty-eight feet higher than the other—before a six hundred-foot section broke off entirely and plunged into Puget Sound.
All bridges are now subjected to full wind-tunnel tests before any concrete is poured or steel fabricated. Some parts of this testing are straightforward—the aerodynamics of the bridge cross-section and its towers are easy enough to model. Where it gets complicated is the introduction into the analysis of the "local wind climate," an incredibly complex study of historic meteorological records, historic wind directions and speeds, and the local topography—are there any accelerating topographical features around that would make normal winds into gales? And what about hurricanes? Not every area is hurricane-prone, but all areas might get hurricanes on rare occasions. How to model for those?
The problem with these local wind climate studies is that there is so much data, far too much to make accurate calculation possible: historic storm intensity data, storm track data, pressure differentials and the like, data on midspan and quarter-point pressures, defections and deflections, cable tensions, and many other variables, some of them with short-term periods and little apparent predictability. Even massive number-crunching computers are not up to the task; and even if they were, it would take far too long to input the data. And so engineers use a curious statistical sleight of hand with a whimsical name to provide answers about the sensitivity of