Impact of complex system behavior on seismic assessment of RC bridges

The damage to reinforced concrete (RC) bridges observed in the 1971 San Fernando, 1994 Northridge, and 1995 Kobe earthquakes prompted significant research efforts within the earthquake engineering community. Strides were made in capturing the true inelastic seismic response of RC bridge piers, and improvements were made to seismic design methods. However, several characteristics of complex RC bridge response are not yet fully understood. The challenge of assessing risk posed to bridges with irregular or curved geometry subjected to multi-directional loading is non-trivial. The effect of the resulting combined interactions on structural response and its implications on system-level bridge vulnerability are assessed in this study. Fragility curves generated for bridges with varying parameters provides a means for assessing the probabilistic impact of these parameters on bridge response. A set of fragility curves capable of representing the true impact of complex geometry, modeling assumptions, and multi-directional system-level response on the vulnerability of RC bridges is developed. These relationships are an improvement over existing curves developed using observational, opinion-based, or uncalibrated numerical approaches. They also avoid inaccuracies that can arise in experimental testing or calibrated numerical analyses due to assumptions made in test set-up, modeling simplifications, and disregard of system-level interaction. Nonlinear time-history analyses are performed on a set of bridge models subjected to a suite of carefully selected seismic records. Fragility relationships are developed from the resulting structural response data. The models are carefully calibrated using a high quality experimental data set from a large-scale hybrid test successfully completed in the Illinois Network for Earthquake Engineering Simulation (NEES) facility. The hybrid simulation of a curved four-span bridge captures the complex interactions generated by combined axial, flexural, shear, and torsional loading. An advanced six degree-of-freedom control scheme and hybrid simulation platform ensures accurate control and full system-level response. Extensive traditional and advanced instrumentation methods provide dense sets of data that can be visualized and processed for assessing structural response and performing model calibration. Curved and straight analytical models are developed using the same set of calibrated model parameters. A suite of records representing a wide array of seismic hazards are applied to these models under varying uni-directional and multi-directional loading conditions. Nonlinear time history analyses are performed for straight and curved bridges with varying 3D loading effects and modeling assumptions. Statistical analysis is performed on the resulting structural response data to generate fragility curves. Variations in these curves represent the individual and combined influence of these parameters on system-level behavior and RC bridge vulnerability. Results further the understanding of complex seismic response of RC bridge systems, while providing an improved set of vulnerability relationships that accurately represent the potential risk posed to these critical components of our infrastructure