Analysis of wood-framed roof failures under realistic hurricane wind loads

Typical North American wood frame houses have trusses that are toe-nailed to the wall top-plates. In order to develop models of roof failures for such structures, considering the effects of wind speed, direction and duration for a storm, as well as the structural details, knowledge of the load sharing and how it is altered by the nail slip at each of the connections is required. The objective of the current paper is to quantify the changes in load sharing during nail/connection slip. This is achieved by simplifying the roof structure to the essential components of a beam, which bends elastically in the direction normal to the trusses, and real, toe-nailed connections. Test results for a single "roof stiffness" and truss spacing show that significant load sharing is not observed until permanent displacements (i.e., damage) to a connection occur. Dynamic load tests, with the same time-varying load applied to each connection, have shown that the load sharing increases with the damage accumulation to each of the connections and changes continuously. Ultimately, this information will be used in the development of analytical models that are able to consider storm duration effects on the performance and failure of wood-frame houses in extreme winds