Performance-Based Earthquake Engineering (PBEE) is computationally demanding, due to the multiple high-fidelity nonlinear dynamic structural response analyses required to compute fragility curves. Local buckling of tubular steel structures is not properly characterized by typical Engineering Demand Parameters (EDPs) such as story drifts or plastic rotation angles. Targeting the two issues above, in this manuscript we propose using state variables based on Lumped Damage Mechanics (LDM) to characterize Local Buckling (LB) in PBEE. Hence, we propose an efficient and innovative procedure for the fragility analysis of complex tubular structures prone to fail due to local buckling. Moreover, local buckling produces a loss of stiffness, with loads transferred to intact or to less-damaged elements. Eventually, the structure forms a global collapse mechanism. Herein, we show how to identify the most likely global collapse mechanism in non-symmetrical tubular structures subjected to random seismic loading. This involves evaluating damage indices in different elements and their correlation, as well as identifying the combination of LB failures that are more likely to form a global collapse mechanism. Fragility curves characterizing the onset of LB at individual elements, and the most likely global collapse mechanism, are constructed. A simple frame structure is addressed, where the accuracy of the LB-LDM model is checked against experimental results. Another case study involving a non-symmetric tubular wharf illustrates the search for the most likely global collapse mechanism, and the derivation of its fragility function.
Read full abstract