Assessment of bridge vulnerability due to seismic excitations considering wave passage effects

• Vulnerability assessment of bridges is carried out under asynchronous ground motions at different supports.
• Non-linear time history incremental dynamic analyses are performed and fragility curves are generated.
• Fragility curves are integrated with hazard to give mean annual frequency of exceeding pre-defined performance levels.
• Various apparent wave propagation velocities and assumed levels of uncertainties in capacities and modeling are considered.
• Bridge response in the two orthogonal directions is investigated.

Uniform ground motion excitation at different supports is typically assumed in practice while performing seismic response analysis of structures with somehow limited footprints. In essence, design of bridges is customarily performed assuming identical signal at all bridge supports. In fact, ground motions may vary at the different supports, especially, for long extended structures, such as long bridges, dams and pipelines. This paper illustrates the impact of the difference in the arrival time of the ground motions on the seismic performance of continuous box girder bridges in both bridge orthogonal directions (longitudinal and transverse). For illustration purposes, a nine-span case-study bridge with a total length of 430 m is considered. Non-linear time history analyses are carried out using opensees software and the “out-of-phase ground motions at different supports” phenomenon is examined using a set of 20 real records originally extracted from the peer (Pacific Earthquake Engineering Center) Strong Motion Database. The analyses are repeated for different apparent wave propagation velocities (namely, 100, 200, 400 m/s, and ∞ designating synchronized arrival time of the signal at all supports) along the bridge longitudinal direction. Results of the non-linear time history analyses performed in an incremental dynamic analysis context are hence manipulated through a probabilistic analysis framework to generate fragility curves associated with various performance levels for the case study bridge. Fragility curves giving the conditional probability of exceeding various performance levels are then integrated with generated hazard curves defining the expected seismic hazard in a specific zone. The outcome of this integration process results in values of mean annual frequency of exceeding pre-defined performance levels.