UC San Diego

The objective of this work is to develop a probabilistic structural seismic performance assessment methodology. To facilitate, guide, and validate this development, the targeted methodology is applied to an actual bridge- foundation-ground system. First, Probabilistic Seismic Hazard Analysis (PSHA) predicts the possible earthquake ground motion intensities (IM) and their probabilities at the structure site based on the seismic environment. The result of PSHA is a seismic hazard curve that represents the mean annual rate (MAR) of IM exceeding any specified value. The probabilistic characterization of the seismic demand in terms of seismic demand hazard curves for several engineering demand parameters (EDPs) is obtained via nonlinear structural seismic response simulations for ensembles of actual earthquake excitations at various hazard levels. A seismic demand hazard curve for a given EDP provides the MAR of the EDP exceeding any specified value and is obtained by convolving the complementary conditional cumulative distribution function of the EDP given IM with the seismic hazard curve. Three potential failure mechanisms, namely pier flexural failure in the lap-spliced region, shear key failure, and unseating, were identified. A set of damage states was defined for each potential failure mechanism to characterize key stages of formation of the mechanism. New and existing predictive capacity models were used to estimate the structural capacity against each damage state. Based on the predictive capacity models and experimental data collected, a fragility function was developed for each damage state, which provides the conditional probability of damage state exceedance given a specified value of the associated EDP. Then for each potential damage state of the bridge, the MAR of damage state exceedance was computed as the convolution of the corresponding fragility function and seismic demand hazard curve. This probabilistic performance assessment methodology was extended to incorporate the randomness in system properties, such as structural and geotechnical material properties. Effects of the randomness in system properties on the bridge seismic reliability were evaluated. Finally, the loss hazard curve of the bridge was obtained through a multilayer Monte Carlo simulation procedure. This curve provides the MAR of the repair/replacement cost of the bridge exceeding any specified value