Research on seismic excitation direction of double-deck curved bridges: A probabilistic method based on the random forest algorithm

Abstract

Double-deck curved bridges can overcome the limitation of space, connecting areas with complex terrains to form a traffic network. However, they have many vulnerable parts and weak seismic performance, which can easily paralyze the traffic network. Seismic fragility can quantify the seismic performance of double-deck curved bridges from the perspective of probability, which provides a reference for the seismic design and evaluation of bridges. Therefore, in this study, the excitation angle of the ground motion was considered a random variable with a uniform distribution to consider its randomness. In addition, based on the random forest algorithm and partial dependence plots, an analysis framework of the multiparameter fragility model of a double-deck curved bridge was established, and an efficient vector model was constructed in the analysis framework to predict the demand and capacity of the double-deck pier. The analysis results indicated that [0, 0.06π] is the most sensitive range of fragility of double-deck curved bridges to seismic excitation angle θ; if θ < 1.56π, the bridge is a typical ductile seismic system owing to the energy dissipation part of the bridge being at the pier; if θ > 1.56π, the damage of the upper constraint system precedes the yielding of the pier, which has a particular role in seismic reduction and isolation; among all possible seismic excitation angles, the unseating risk of the upper beam is higher than that of the lower beam. The spacing between the upper beam and its limit device can be appropriately increased or double-layer limit devices can be set to enhance the seismic performance of double-deck structures.