Refined Probabilistic Seismic Demand Modeling of High-Speed Railway Bridge Under Near-Fault Hanging-Wall Effects and Velocity Pulse Effects by a Novel Ground Motion Selection Approach

Near-fault ground motions may cause serious damage and consequence of high-speed railway bridges (HSRBs), since their particularly detrimental effects, including the hanging-wall effects and velocity pulse effects. However, the induced seismic demand depends on the complex relationship between natural vibration characteristics of structures and frequent contents of ground motions, leading to the difficulty in attributing to less seismic attenuation or/and the above special effects. This work adopts a framework to select a ground motion suite with a specified near-fault effect by multiple controllable constraints on the target response spectrum. Near-fault effect on seismic demand is visualized and quantified by empirical cumulative distribution functions (ECDFs) and the Kolmogorov-Smirnov test. An ECDF-based method is proposed to develop a bilinear probabilistic seismic demand model (PSDM) to capture the nonlinear relationship between engineering demand parameter (EDP) and intensity measure caused by near-fault effects. In the case study, the hanging-wall effects and the directivity-induced velocity pulse effects are investigated on a simply-supported HSRB. The influence of the two effects on key structural components is examined, in which the demand changes are attributed to one of them. The result shows that the hanging-wall effect amplifies both mean and variance of the seismic demands. Pulse-like records cause sudden increases in demand ECDFs, which are used to determine the transition point in the bilinear PSDMs. The refined PSDMs well capture the nonlinear trend of EDP varying with PGV when considering pulse-like records.