Seismic Risk Analysis of Bridge Structural Systems.

The objective of this Ph.D. Dissertation is to develop a new model for the seismic risk assessment of bridge structural systems. The model is designed to account for the random nature of the seismic input, the variability in the material properties, and also for the modeling uncertainties or systemic uncertainties that reflect our lack of knowledge of all the factors that control the seismic response of bridges and the ability of the bridge structural system to withstand seismic ground motions. To accommodate the large number of random variables and the various modes of failure that may exist in a bridge system subjected to seismic input, this study develops a new reliability method herein called the Radial Sampling method to calculate the probability of failure of each mode of failure as well the overall system's probability of failure. The application of the proposed risk analysis methodology to study the reliability of reinforced concrete bridge columns has led to proposing a rational approach for selecting appropriate response modification factors, normally used for the force-based design of typical bridge configurations in low seismic regions, that would lead to uniform and consistent levels of safety. The efficiency of the proposed risk analysis approach is tested on a nonlinear finite element analysis model of a typical bridge configuration. The material varaibilities are studied and identified from existing experimental data on the dynamic response of reinforced concrete columns. The resulting material statistical models and the seismic uncertainties defined based on the seismic hazard analysis are introduced into a nonlinear dynamic reliability analysis model that is solved using the proposed Radial Sampling method. The example demonstrates the ability of the Radial sampling method to efficiently handle large numbers random variables and identify the most important failure modes as well as calculate each mode's probability of failure and the failure probability of the complete structural system. The models developed in this study can be used to set up a preparedness plan, following the occurrence of an earthquake event that speeds up the recovery process by identifying and targeting for repair high seismic risk members of critical bridges. The models can also serve to develop a bridge rehabilitation strategy that rationally uses available resources to optimize the chances of survivability and provide a uniform and acceptable seismic risk.