A few months ago, I discussed an example of an isolated wing of an old steel-frame building being connected back to the main block. It’s worth discussing the way this use of struts changes the analysis of such a building.
The picture above is the back of a 1920s apartment house on west 23rd Street, exposed by a 1960s reconfiguration of 24th Street that removed the north half of the block. The plan of the building is roughly a capital E, with a solid front facing south on 23rd and three equal wings, separated by two deep light courts, facing north toward 24th.
A critical point in understanding these struts is to remember that the steel frame of this building was designed for gravity and wind loads but not seismic load. I’ll go through the wind-only lateral load version first and then the seismic-load version. The difference is simple: wind is an external force acting on the building that comes from a specific direction.
If the wind is blowing from the north or south, the light-courts have little effect on the analysis, since there are a series of internal frames, designed to resist the wind, made up of the columns and girders connecting them in the north-south direction.
If the wind is blowing from the east – the left side of the picture – the full force of the wind pressure is taken by the eastmost wing, with a lesser but still significant amount of wind suction taken by the westmost wing. The center wing would be effectively unloaded by the wind. At the very least, this is obviously wasting the potential capacity of one third of the structure to resist wind, but the problem is actually worse than that. A wider structure resists wind better: a wing one-third the width of the whole building will be much less than one-third as strong and stiff than the building as a whole unless its interior structure is made much heavier. The E layout robs the building of most of its east-west resistance to wind.
The struts could be part of the northmost wind-resisting frame at the 12th floor, working in bending to contribute their strength to the beams and columns in the adjacent windward wing, but that’s not their only contribution. They also act as compression or tension struts connecting the wings to one another. If we picture the east wing bending to the south as wind pushes from the east, and the west wing bending a bit less to the south as the wind suction pulls from the west, the struts will force the center wing to move along as well. In short, if the struts force the three wings to move sideways the same amount, then they force the northmost frames of all three to work equally hard. The result of these two effects is not as strong or stiff as if the light courts dod not exist, but it’s stronger and stiffer than the sum of the strength and stiffness of the three wings and therefore significantly stronger and stiffer than the sum of the strength and stiffness of the two wings that are loaded by wind.
This type of strut is of much less use for earthquake design because seismic loads are not external, but rather come from the weight of the building itself acting as the earth moves below it. So an earthquake loads all three wings equally to begin with, and sharing load between them does not improve the situation. If the original designers were addressing earthquake loads, they would probably have used a different solution.