Simulation-based seismic reliability assessment of complex structural systems

Researchers from various institutions have recently developed innovative steel moment frame systems for earthquake-resistant design that, following a major earthquake, have the potential to reduce or eliminate structural damage and return to its original vertical position (i. e. self-center). As the development of these steel moment frames with self-centering behavior progresses, it will be necessary to specify design parameters and a workable design method. This will require a good understanding of the sensitivity of the structure to changes in these parameters as well as the reliability of structures built based on the proposed method. This research investigates a risk assessment methodology (using Monte Carlo simulation) related to the seismic response of self-centering steel moment frame prototype buildings. A complex and a simplified model for a Self-Centering Steel Moment Resisting Frame (SC-MRF) are developed and validated against past data. Next, a method of reliability assessment is proposed. A series of conditional seismic reliability assessments are performed on a structural model. Two methods are developed to generate synthetic earthquakes: one using a specific building site which uses seismic hazard data to generate events with realistic magnitudes and distances based on a return period earthquake (site characteristic method), and another generating events based on a scenario (scenario method). The earthquakes are applied to the structural model and data is recorded as to the structure's response. This data is then used to generate demand curves (with respect to the structural response of interest). Next a capacity curve is generated for a structural limit state. The results of both studies are combined to yield the overall limit-state specific reliability of the structure. To illustrate the effectiveness of the method, it is applied to the simplified model of the SC-MRF model developed. An example is presented for both the site characteristic method of event generation, and the scenario method. In addition, a simple parametric study involving hysteretic energy is performed to illustrate parametric sensitivity analysis applied to the reliability assessment. Overall the research performed in this dissertation was found to provide an effective method of seismic reliability assessment of complex structures.