Manmade structures can be described as a collection of smaller substructures joined together at contacting interfaces. Contact nonlinearities present at these interfaces can cause computational bottlenecks in large, high-fidelity finite element models. Consequently, different techniques of model order reduction have emerged, wherein the number of degrees of freedom (DOF), particularly at nonlinear interfaces, is decreased to a manageable level. In the analysis of civil structures, so-called “macro” elements are often employed, which represent entire contact surfaces using uniaxial spring and dashpot elements. Mechanical and aerospace systems, on the other hand, more often use projection-based model order reduction, which transform interface deformations onto a low-dimensional subspace. This research develops high-fidelity and reduced-order models of contacting structures, from large-scale building structures down to small-scale mechanical components. Moat wall pounding, a phenomenon where base-isolated buildings collide with their retaining walls during seismic shaking, is examined as a case study for reduced-order models of civil structures. In the first half of the dissertation, new and existing macro impact elements are verified against experimental data, and then statistically examined in a large parametric study. Following that, two different high-fidelity models of moat wall pounding are developed and compared with corresponding experimentation. Results from these studies indicate that most macro models provide sufficient accuracy for moat wall pounding simulations. Their high-fidelity counterparts, while more difficult to calibrate to experimental data, reveal dynamic behavior that cannot be ascertained with macro elements alone. The second half of the dissertation develops a novel family of methods to reduce the nonlinear interface DOF for preloaded bolted structures. Five such methods are applied to a bolted beam assembly and compared in terms of accuracy and computational savings. Results show that the interface reduction methodology is capable of capturing often-neglected interface kinematics, including transient contact area and joint slip. Furthermore, the methods are tunable to the desired combination of accuracy and computational effort.

Seismic isolation can be an effective strategy to protect critical facilities including Nuclear Power Plants (NPPs) from the damaging effects of horizontal earthquake ground shaking. For critical facilities, the isolation system should demonstrate a high-confidence of low-probability of failure at the design level and the load carrying capacities should be maintained under beyond design earthquake shaking (BDBE). Experimental evaluation of seismic isolation bearings is important to fully understand their behavior and capacity for reliable performance. Safety mechanisms such as a stop can be imposed to prevent excessive displacement of the isolation system under BDBE, however, this raises concerns for detrimental effects of pounding against a stop or moat wall. Methods of analysis are presented in this dissertation to evaluate both seismic isolation system behavior under extreme earthquakes and the potential effects of pounding by imposing displacement restraints.

The dynamic response of an isolated NPP depends on the combined characteristics of the ground motion, bearings, and structure while the seismic isolation bearings themselves can exhibit complex nonlinear behavior that depends on several factors, including the scale size, axial load, temperature, and rate of loading especially under strong earthquake shaking. With a specific interest on the in-structure response of seismically isolated NPPs, hybrid simulation is shown to be a viable approach to examine bearing behavior at full scale under realistic earthquake loading. The adaptation of a full-scale bearing test machine (SRMD testing facility at UC San Diego) and developed toolsets for the implementation of fast hybrid simulation to study the dynamic response of base isolated NPP using full scale lead plug rubber bearings under realistic earthquake loading conditions are presented. Results from these tests validate the effectiveness of seismic isolation technology for application in nuclear facilities and provide valuable data towards improving numerical models of seismic isolation bearings.

In a seismically isolated NPP, a surrounding moat wall can function as a stop to limit isolation system displacements and prevent bearing failure for beyond design basis shaking. Impact of isolated structures against a moat wall is of concern due to potential amplification of superstructure response. A moat wall model able to capture impact forces is proposed and used in numerical simulations to capture the effects of impact on the response of seismically isolated NPPs. Variable clearance to the stop and a range of properties for the impact model, moat wall and isolation system are considered to identify parameters that influence the response. Results indicate that large NPP plants as considered here can have significant penetration into the moat wall, not fully limiting displacements in the isolation system, while causing considerable increases in accelerations throughout the NPP. A simplified methodology to estimate impact response parameters including penetration is proposed towards developing design tools that consider these effects.

Seismic isolation is an established effective strategy to protect building structures from earthquake-induced damage by reducing accelerations and drifts within the superstructure. Current design codes primarily focus on ensuring operational resilience under design-level earthquakes with the building structure designed to remain essentially elastic. However, there is limited guidance for assessing the nonlinear behavior of the superstructure for beyond-design seismic events, which can be of concern since the structural performance can degrade rapidly from this point. This research investigates the performance of seismically isolated structures under extreme loading conditions, with emphasis on the response of the superstructure. The motivation for this research is to expand rapid assessment tool for the seismic performance evaluation of seismically isolated buildings. This expansion aims to enhance the understanding of limit state behaviors for beyond-design basis events.Utilizing a two-degree-of-freedom (2DOF) model, this study explores the dynamic interactions between the isolation system and the superstructure, highlighting the limitations and reliability of simplified models compared to full nonlinear analyses. Through Incremental Dynamic Analysis (IDA), the study evaluates various archetype buildings with different lateral force-resisting systems, examining the effects of parameters such as yield strength, ductility capacity, strength degradation, and moat wall pounding on the overall performance. The results indicate that while 2DOF models can provide a rapid assessment of the isolation system lateral displacements and superstructure drifts, they tend to underestimate roof accelerations and overestimate interstory drifts. Furthermore, the occurrence of moat wall impacts is shown to significantly affect the structural response, emphasizing the critical need for design strategies that mitigate or avoid such impacts. The research contributes to the broader understanding of seismic isolation, offering valuable insights into the performance of isolation systems under extreme seismic loading conditions.

Seismic isolation is an effective method to mitigate the damaging effects of horizontal ground motions. The flexible layer, typically placed at the base of the structure, reduces forces transmitted from ground shaking at the expense of concentrated displacements at the seismic isolation layer. Base isolation systems within a basement of a building require a free clearance to allow for such displacements to occur. A surrounding moat wall can be placed to constrain the isolation devices from exceeding their displacement capacity. However, impact to the moat wall can be damaging to the structure with recent studies concluding that the required clearance to stop (CS) specified by building design codes is insufficient for high consequence low-frequency ground motions.The Lead Rubber Bearing (LRB) is widely used in practice for implementation of seismic isolation. Current models for LRB are not able to capture the salient characteristics of bearing behaviors observed in experimental data, especially under large displacement demands. Therefore, the large strain lead rubber bearing (LSLRB) model is proposed to better predict the response of base isolated structures under extreme earthquake shaking considering the combined effects of lead core heating, and material strain hardening in the lead and rubber. The LSRLB model was implemented in a full-scale Nuclear Power Plant (NPP) numerical model under earthquake loading and demonstrated to reduce displacement demands and lower velocities in the case of impact to a moat wall. Consideration of extending the moat wall clearance and allowing the bearings to reach strain hardening at large displacements showed to be effective in improving the overall seismic response under large ground motions. In terms of the effect of bearing models on the critical internal contents of NPPs, the LSLRB model showed a reduction in floor spectral accelerations throughout the superstructure compared to current models utilized in practice. The results presented are also highly dependent on modeling of the moat wall impact. Current models are reviewed and extended to better predict the amount of moat wall deformation considering the concrete retaining wall, soil contribution and coefficient of restitution.

Steel moment frames are commonly used in seismic regions, with wide flange steel columns serving as essential elements in their design. Compliance with drift limit requirements updated after the 1994 Northridge earthquake has led to the frequent selection of deep and slender columns, which maximizes the moment of inertia of the section for more economical design. However, such columns are susceptible to experiencing local and global buckling and subsequent axial shortening when subjected to high axial forces and cyclic lateral loads. This study presents the development and experimental observations of a full-scale testing program of cruciform beam-to-column subassemblies through quasi-static and advanced hybrid simulation.A novel framework for hybrid simulation is presented that employs a mixed displacement and force control strategy. A key feature of this framework is a controller-based displacement-to-force transformation that enables compatible displacements between the numerical and experimental model for force-controlled degrees of freedom. The framework can be applied within conventional displacement-based time-integration algorithms allowing for implementation across a variety of software and hardware architectures; it is demonstrated here using OpenSees software with detailed nonlinear numerical models. This study includes numerical verification of the proposed force control strategy and the later implementation for the hybrid simulations as part of the experimental program. In the hybrid tests, the experimental column axial load is applied in force control due to the large stiffness and when combined with lateral seismic loads, result in column buckling with significant axial shortening. The hybrid simulation results verify that the proposed mixed displacement and force control strategy effectively enforces displacement compatibility and force equilibrium between the numerical and experimental structures at the boundary degrees of freedom. Additionally, hybrid tests include an online model updating algorithm focused on the modeling parameters of plastic hinge elements representing the reduced beam sections in the numerical model. A smooth plasticity model is utilized for beam plastic hinges with updating parameters identified from on-line experimental data through a modified version of the unscented Kalman filter. The hybrid simulation shows that the numerical beam hinges based on simple hysteretic model with updated parameters are able to capture the characteristic behavior observed in experiments even for cases involving fracture. A selective updating concept is proposed to allow for updating multiple numerical component accounting for asymmetric behavior and variability in the response. The selective updating method is validated through virtual hybrid simulations that are better able to identify and isolate the effects from different simulations. The combination of results from physical and virtual tests highlights the benefits of model updating on the local and overall system-level response. Data from the present study and previous tests were employed to develop a column hinge macro-element. It allows for simulating the moment-rotation response of deep columns through fiber-based elements, while the vertical direction is modeled via an empirical expression based on the measured axial shortening. The macro-element is included in a steel moment frame model to assess the axial shortening for different seismic intensities.