Understanding, predicting, and preventing collapse have always been major objectives of earthquake engineering. Collapse is the main source of injuries and loss of lives. Thus, it constitutes an engineering limit state that needs to be predicted in order to evaluate, in a probabilistic format, the life safety performance level, which is of primary societal concern. In the context of earthquake risk management, a process is needed that permits a rigorous assessment of the probability of collapse to make informed decisions in the best interest of society. In the context of earthquake resistant design, this process needs to be simplified so that the engineering profession can use engineering techniques, which are based on parameters such as strength, stiffness, and ductility (deformability), to derive structural properties that comply with specified targets for a required level of collapse safety, or a tolerable probability of collapse. This project will address both contexts. It will provide a methodology and reliable data for predicting a critical mode of collapse, namely that associated with sidesway instability in which an individual story (or a series of stories) displaces sufficiently so that the second order P-delta effects fully offset the first order story shear resistance and dynamic instability occurs, i.e., the structural system loses its gravity load resistance. Prediction of this mode of collapse is a challenging problem because structural components will deteriorate in strength and stiffness before the collapse limit state is reached, and great uncertainties are associated with the description of the seismic input and of the parameters that control the response of structures close to collapse. The methodology will be based on a combination of analytical and experimental simulations, with the former being carried out at Stanford University and the main effort of the latter, a shaking table collapse test of a model of a steel structure, being carried out at the NEES facility at the University at Buffalo. The outcomes of the proposed research will be (a) a methodology for predicting sidesway collapse of deteriorating structural systems, (b) an extensive database on deterioration properties of structural steel and reinforced concrete components, including uncertainty measures accounting for modeling and material uncertainties, (c) incorporation of component deterioration models in the OpenSees platform, (d) documentation of a comprehensive collapse experiment with data that covers the range of response of a 5- story steel frame structure from elastic behavior to incipient collapse, (e) a methodology for computing the probability of sidesway collapse that accounts for hazard, ground motion, and structural (material, modeling) uncertainties, and (f) engineering recommendations for design for collapse safety. Intellectural Merit. The proposed research will lead to seminal advances in understanding and predicting sidesway collapse of structures subjected to severe earthquakes. The major challenges are to account for deterioration in structural behavior in combination with P-delta effect, and for the uncertainties inherent in ground motion description and in modeling the behavior of the components that control the dynamic response at large inelastic deformations (or at small deformations in the case of brittle elements). These challenges imply extensive modeling efforts at the component and structure level and the incorporation of reliability concepts in assessing collapse safety of structures. Another major challenge will be the planning and execution of a shaking table collapse experiment, whose main purpose is to demonstrate that collapse prediction indeed is feasible. The research involves an integrated approach of physical simulations at a NEES site shaking table and computational simulations on the OpenSees platform, while fully utilizing the simulation, visualization, and collaboration tools of the NEESgrid. Broader Impacts on Earthquake Engineering Research and Practice. The research is motivated by professional needs to provide adequate and consistent collapse safety in new designs and to assess the collapse hazard of existing structures. These needs require the filling of knowledge gaps that exist in regard to data and tools and the understanding of collapse phenomena. The impact of the proposed work on research and practice will be the development of databases, advanced deterioration models, and computational tools that will make it feasible to predict the collapse safety of complex structures and will permit more rational allocation of resources in the context of performance-based seismic risk management. The impact will be felt in the academic environment in teaching and research, in engineering design offices, and in organizations concerned with risk management.
