Strength and Reliability of Structural Steel Roofs Subjected to Ponding Loads

Abstract

Roof collapses due to ponding instability are among the most frequent structural failures. Although methods of design for other types of instability have evolved and improved with computational advances, the most commonly used method of assessing ponding instability has been updated minimally since it was developed over 50 years ago. Newer methods of design for ponding have been proposed, but they lack independent verification of their ability to assess the strength of roofs and to provide a sufficient level of reliability. An advanced analysis approach capable of capturing material and geometric nonlinearity, as well as accumulation of water caused by deflection, was developed for roofs consisting of steel beams on stiff supports. The new analysis approach generates data against which the various methods of design for ponding can be benchmarked. These comparisons showed that methods of design based on elastic analysis are capable of accurately capturing the strength of structural steel roofs subjected to ponding load if a reduction factor of 0.8 is applied to the stiffness of the structure. In addition, a reliability analysis using Monte Carlo simulation showed that the methods of design for ponding can achieve a target reliability index consistent with design for other loading conditions. The results provide new insights into the behavior of roofs subjected to ponding loads and enable the use of new design methods for ponding that are more broadly applicable, comprehensive, and consistent with current methods of strength evaluation.