Optimal intensity measures and probabilistic fragility assessment for the long-span aqueduct structure with four-column bents

The intensity measure (IM) is a key factor in determining the accuracy of seismic performance assessment, which can represent the power of ground motion. However, the optimal IMs selection for long-span aqueduct structures with four-column bents has not been investigated in the previous study. This study aims to evaluate the optimal IMs for use in probabilistic seismic demand model (PSDM) and propose the scalar- and vector-valued fragility methods of the long-span aqueduct structure. To achieve this goal, taking the long-span aqueduct structure with four-column bents in the Central Yunnan Water Diversion Project in Southwest China as a typical case, a series of nonlinear dynamic time history analysis are conducted. Then, the 21 commonly-used IMs are tested and evaluated based on the different metrics (e.g. correlation, efficiency, practicality and proficiency). The optimal vector-valued IMs for the fragility analysis of the long-span aqueduct structure with four-column bents is proposed. Meanwhile, the PSDMs of optimal vector-valued IMs are also established and compared. Finally, the scalar- and vector-valued fragility functions are developed in terms of the optimal IMs. A method for seismic fragility analysis of the long-span aqueduct structure with four-column bents based on vector-valued IMs is proposed. The numerical results reveal that peak ground acceleration (PGA), peak pseudo acceleration spectrum (PSA), root-mean-square of acceleration (A_rms) and characteristic intensity (I_c) are relatively appropriate IMs for seismic performance evaluation of the long-span aqueduct structure with four-column bents. In particular, PGA is considered to be the optimal IM due to its highest correlation, better efficiency, practicality and proficiency. The scalar-valued fragility curves can only describe the impact of the single IM on the damage probability of the aqueduct structure, which may overestimate or underestimate its damage probability. Furthermore, the vector-valued IMs can significantly increase the fitting ability of the PSDMs. The fragility surfaces are superior to scalar-valued fragility curve due to the vector-valued IMs can more accurately describe the ground motion information. In summary, the findings of this study highlight the significance of proposing the optimal IMs and developing vector-valued fragility surfaces when evaluating the seismic performance of the long-span aqueduct structure with four-column bents.