Reliability and performance-based design, assessment, and rehabilitation for RC structures located near active faults

In this dissertation probabilistic models and procedures are developed for performance-based seismic design of new buildings, seismic vulnerability assessment of existing buildings, and rehabilitation and upgrading of hazardous buildings. The study focuses on three main components of seismic design/assessment: the critical dynamic characteristics of the earthquake ground motions (EQGMs) regarding the damage potential for different performance levels; the structural and nonstructural demands; and the structural capacity. In this study, only two performance levels are considered: the fully-operational performance level (FOPL) as representative of elastic response and the life-safety performance level (LSPL) as representative of inelastic response of structures.; Although random-type EQGMs from near and far sources are considered, emphasis has been placed on near-source ground motions characterized by severe pulses, because the elastic and inelastic responses of structures to such motions are different from those to random-type EQGMs. Special attention has been devoted to the drift ratio response because this is one of the most important responses in evaluating damage. The shortcomings of some existing methods for estimating elastic response of structures are identified and improved solutions for estimating the drift ratio response of damped shear beams are presented which are based on wave propagation and frequency domain analyses.; Two new measures of ground motion severity are introduced to select critical EQGMs and quantify their severity. These measures are the velocity equivalent peak ground acceleration (VEPGA) and the significant peak ground acceleration (SPGA). It is shown that these two measures are strongly correlated with the inelastic response of short period structures.; The Bayesian parameter estimation technique is used to develop probabilistic capacity and demand models for LSPL. For capacity models, mechanics of the behavior along with experimental data are used and significant improvement in predicting capacities is achieved. For demand models simulated data from nonlinear dynamic analyses of structures are used. The probabilistic capacity and demand models are used to estimate the fragility of a RC structural wall and a RC frame system for given SPGA at the LSPL. The study concludes by proposed performance-based design, assessment and rehabilitation procedures, based on a reliability approach.