Hello Mike;

I don't have a copy of D1.1 at hand here  in my hotel room, however, check section 2, Design. Look at the chart that shows the S-N curves for different fatigue categories. I may be wrong, but I believe the S-N curves kick into play only if you exceed the number of cycles listed to the extream left of the table. The horizontal axis is a logrithmic scale of loading cycles while the vertical axis is the stress.  If the anticipated number of cycles is less than the value along the left end of the horizontal axis, fatigue isn't a problem. Likewise, the load has to be above a certain threshold inorder to be a consideration.

The point being, if the unit stress is very low or if the number of loading/unloading cycles isn't above certain threshold values as indicated by the S-N chart, fatigue isn't a concern. 

I believe your example of the wind blowing against the side of a building is exactly what I am talking about. A light breeze isn't going to develop enough load to produce a unit stress that will be of any consequence. However, a hurricane force wind will induce a unit stress that is of a magnitude to be of concern if the building experiences a sufficient number of hurricanes in it's expected lifetime.

Fatigue life is an inverse function of the unit stress versus the number of load cycles experienced. As the unit stress increases, the number of cycles necessary to cause fatigue is reduced. Hence, the S-N curve rises toward the left side of the chart (increased unit stress) and slopes downward (decreased unit stress) as you increase the number of loading/unloading cycles. If the unit stress is "low" enough, the number of loading/unloading cycles becomes infinate (i.e., the S-N curve is flat and parallel to the horizontal axis).

Best regards - Al

Hello Mike;

I'm back in my office and I just took a look at the S-N chart in D1.1. The chart, figure 2.11, does start at 10,000 cycles. For the worst case, i.e., the most suseptable geometry, provided the level of unit stress is below 2.5 ksi, the endurance limit is infinate (100 million cycles). So, I would expect that fatigue has to be considered if the expected cycles exceeds 10,000 and the unit stress is in excess of 12 ksi for the worst case geometry. If the geometry is improved, the unit stress can increase as well as the number of cycles.

Best regards - Al