Numerical and experimental methods for seismic risk assessment of civil and industrial structures

Due to high seismic vulnerability and severity of possible failure consequences, petrochemical installations are often considered as “special risk” plants. Although tanks, pipes, elbows and bolted flanges have been a major concern in terms of seismic design, generally, they have not been analysed with modern performance-based procedures. This thesis will explore some important themes in seismic risk assessment with a special focus on petrochemical plants and components. In the first part of the thesis the case study of a probabilistic seismic demand analysis (PSDA) for a Refrigerated liquefied gas (RLG) subplant is presented. As a matter of fact, RLG terminals that are part of strategic facilities must be able to withstand extreme earthquakes. In detail, a liquefied natural gas (LNG, ethylene) terminal consists of a series of process facilities connected by pipelines of various sizes. In this study, the seismic performance of pipes, elbows and bolted flanges is assessed, and seismic fragility functions are presented within the performance-based earthquake engineering framework. Particular attention is paid to component resistance to leakage and loss of containment (LoC) even though several different limit states are investigated. The LNG tank, support structures and pipework, including elbows and flanges, are analysed with a detailed 3D finite element model. For this purpose, a mechanical model of bolted flange joints is developed, able to predict the leakage limit state, based on experimental data. A significant effort is also devoted to identification of a leakage limit state for piping elbows, and the level of hoop plastic strain was found to be an indicator. The second part of the thesis describes an innovative methodology to evaluate seismic performances of a realistic tank-piping system with special focus on LoC from piping elbows. This methodology relies on a set of experimental dynamic tests performed throughout hybrid simulations where the steel storage tank is numerically modelled while, conversely, the physical substructure encompasses the coupled piping network. Besides, ground motions for dynamic tests are synthetized based on a stochastic ground motion model whose input parameters are derived from the results provided by a seismic hazard analysis. Then, based on output data from the experimental tests, both a high-fidelity and a low-fidelity FE model are calibrated. Furthermore, these models are used to run additional seismic analyses using a large set of synthetic ground motions. Moreover, in order to derive the seismic response directly from inputs parameters of the stochastic ground motions model, the procedure to build a hierarchical kriging surrogate model of the tank-piping system is presented. Eventually, the surrogate model can be adopted to perform a seismic fragility analysis. Along with the line of probabilistic analysis, another contribution to this research work is a probabilistic seismic demand model (PSDM) of a steel-concrete composite structure made of a novel type of high-strength steel moment resisting frame. According to the main topic of this thesis, the procedure that is here presented can be used either in a seismic risk assessment or a fully probabilistic performance-based earthquake engineering (PBEE) framework. In detail a 3D probabilistic seismic demand analysis was performed considering the variability of the earthquake incident angle, generally not taken in account in typical fragility analyses. Therefore, the fragility curves evaluated following this approach account for the uncertainty of both the seismic action and its direction.