Multiscale Simulation and Assessment of the Seismic Resilience of Communities

Abstract

Quantifying the seismic resilience of communities requires rigorous modeling of their behavior at disparate temporal (earthquakes – seconds vs. recovery – months) and spatial (component - meters vs. system - kilometers) scales. Hence, this dissertation has two main goals. The first one is to investigate the seismic behavior of components with heterogeneous scales in the community (i.e., member, building and community level studies) and further explore the effect of their behavior on the seismic resilience of communities over the relevant time scales. The second goal is to investigate the mutual interdependencies between the different systems of the community (i.e., engineering, social, etc.) during the disaster and the post-disaster recovery stages. On the member level, measurements obtained from a 3D noncontact laser scanning technique are used to quantify the initial geometric imperfections of steel W-shape members. Based on the measured imperfections, a spectral approach that models the imperfections in each plate of the W-shape member as a 2D field of random vibrations is proposed. It is shown that although geometric imperfections can, in certain situations, influence column buckling behavior, their effect on nonlinear cyclic behavior is generally small and inconsistent. The capabilities of different machine learning classification and regression methods in predicting the seismic collapse behavior of deep steel W-shape columns in SMFs are explored. A dataset of more than nine hundred experimental and numerical results of deep steel W-shape columns with different attributes is assembled. The results suggest that machine learning algorithms that are continually updated with new experimental and computational data could inform future generations of design specifications. The seismic collapse behavior of SMF hollow structural steel (HSS) columns under combined axial and drift loading is computationally studied through a validated finite element model. The simulation results are used to propose slenderness limits and design guidelines that incorporate key variables identified in the research to permit HSS columns to achieve highly ductile behavior. On the building level, the extent of debris generation around collapsed reinforced concrete moment frame (RCMRF) buildings is characterized using a validated computational approach. A set of RC moment resisting frame structures with different heights is modeled under different ground motion records scaled up until they induce collapse of the building to assess the seismic debris field under different ground motion histories and building heights. The effect of building code requirements on debris field extent is also investigated. On the community level, a scalable model that employs a simulation-based dynamic analysis, which models the behavior of the community at each time step as the seismic event occurs (time step in seconds) and as the community recovers after the event (time step in days) is developed. The developed model is employed to simulate the mutual interdependencies between the building portfolio, transportation network, and healthcare system in the community as well as to integrate post-earthquake household decision making when quantifying the seismic resilience of communities subjected to earthquake sequences. Incremental dynamic analysis (IDA) is used to develop fragility curves for mainshock-damaged structures, which are distinguished from the conventional fragility curves of intact structures. The capabilities of the developed models to support hazard mitigation planning are demonstrated through various case studies that highlight the effects of interdependencies between the different systems under consideration. Mitigation strategies to improve seismic resilience of the prototype communities are also proposed and assessed.