The sufficiency criterion has long been used to evaluate the effectiveness of a ground motion intensity measure (IM) in capturing the link between ground shaking and structural response. However, a typical sufficiency-based evaluation of an IM only tests for the possibility of linear dependency and the interaction among the upstream parameters is not considered. To address these and other limitations, two new IM evaluation methodologies are proposed. The first methodology considers the loss of statistical information when an IM is used to predict the engineering demand parameters (EDPs) without including the upstream parameters (i.e., earthquake magnitude, source-to-site distance and epsilon). The best IM is the one that minimizes the loss of predictive performance when it is the only model input relative to when it is used as a predictor together with the upstream parameters. To consider the possible interactive effects, a machine learning model is used when both the IM and upstream parameters are used as inputs. The second methodology uses a causal inference approach where the effect of the IM on the EDP distribution is quantified while considering the earthquake magnitude, source-to-site distance and epsilon as control variables. The double machine learning approach is implemented for this purpose. The two methodologies are applied to a set of five steel specifical moment resisting frames. The results show that the statistical loss-based and causal inferencing approaches produce results that are more conclusive than the sufficiency-based approach and more consistent with the physical laws that govern the IM-EDP relationship.

Post-earthquake damage and recovery assessment of individual and portfolios of building are critical aspects of earthquake risk reduction and enhancing resilience. The application of such predictive models is a necessary step in quantifying the social and economic consequences of an earthquake. However, the myriad of processes and influencing factors that these models attempt to capture are quite complex with many interactions. As such, the collection and use of empirical data are critical for their successful development and implementation. This research presents a comprehensive study on the collection and curation of relevant data and the use of Bayesian inferencing to model post-earthquake damage and recovery of buildings. A dataset of observed building damage and recovery following the 2014 South Napa earthquake is assembled on a publicly available relational database. The information categories that comprise the dataset include the building geometric and structural characteristics, site properties and ground shaking intensities, documentation of the damage, permitting and repair (descriptive and duration-based) of the affected inventory and census-block-level sociodemographics. The first-of-its-kind database provides a blueprint for organizing information that is collected in the wake of a natural hazard event. A Bayesian parameter estimation methodology is then developed to update the distributions of the duration-based variables used in post-earthquake building recovery modeling. The formulation recognizes the high degree of uncertainty that is inherent in modeling complex recovery processes and provides a framework for balancing prior expert knowledge and newly available data sets. A methodology for dynamically updating of building damage and recovery is also developed. The goal here is a provide the methodological underpinnings for making use of the continuous in flow of information as the recovery unfolds over time. Lastly, Bayesian parameter estimation of building damage fragility functions are implemented using a subset of the Napa data set. The fragility parameters provided in HAZUS are used as the priors and the updating is performed using both the conjugate prior and Markov Chain Monte Carlo methods.

With the embrace of the performance-based seismic design as the state-of-the-art design method, recent emphasis has been placed on eliminating its drawbacks and facilitating its application in practice. This study aims to propose an alternative design method: performance-based analytics-driven seismic design, which is applied to steel moment resisting frame buildings. First, the seismic performance of self-centering (with post-tensioned connections) and conventional moment resisting frames (with reduced-beam section connection) is comparatively assessed. The comparison indicates that the economic benefit for adopting the post-tensioned connection is not significant. Then, an end-to-end computational platform, which automates the seismic design, nonlinear structural model construction, and response simulation (static and dynamic) of steel moment resisting frames is developed. Using this platform, a comprehensive database is developed, which includes 621 special steel moment resisting frames designed in accordance with modern codes and standards and their corresponding nonlinear structural models and seismic responses (i.e., peak story drifts, peak floor accelerations, and residual story drifts). Using this database, the efficacy of mechanics-based, data-driven, and hybrid (combination of mechanics-based and data driven) approaches to estimating the seismic drift demand are evaluated. The evaluation results reveal that the hybrid approach has the best performance whereas the mechanics-based model has the lowest performance. Next, a set of non-parametric and parametric surrogate models are developed for estimating the engineering demand parameter distributions. A comparative assessment of the proposed surrogate models and the simplified analysis method proposed by FEMA P-58 is conducted to demonstrate the superior predictive performance of the former. Finally, the effect of various design variables on the collapse performance of steel moment resisting frames are evaluated. The research findings presented in this study helps to facilitate the application of 2nd performance-based earthquake engineering framework in practice and thus better help to create earthquake-resilient communities.

Seismic resilience of infrastructure has received growing attention in the past decades. Thisdissertation aims to propose a general decision support framework for the post-earthquake damage and recovery of distributed infrastructures in order to improve their resilience. A damage assessment framework is first proposed that incorporates a neural network-based pre-event assessment model, and a graph neural network-based model for dynamically updating the damage estimates in the recovery process. Next, a spatially explicit model is developed to quantify and estimate the possible recovery trajectory based on the Gaussian Process model. Finally, a deep reinforcement learning based framework is proposed to optimize the repair actions in the recovery process. The framework can provide critical insights to the recovery of distributed infrastructure and improve pre-event planning and post-earthquake recovery tasks.

This study is focused on assessing the seismic resilience of two 42-story reinforced concrete dual system tall buildings designed by different methods. A systematic rating approach based on the United States Resiliency Council (USRC) Seismic Rating procedure is used as the resilience metric. The two buildings were developed as part of the Pacific Earthquake Engineering Research (PEER) Institute’s Tall Building Initiative (TBI) project. Both buildings were designed based on an assumed site location in Los Angeles, California. One variant was designed using prescriptive code provisions and the second using the Los Angeles Tall Building Structural Design Council (LATBSDC, 2008) Guidelines. The Seismic Performance Prediction Program (SP3) is used to perform the building rating analysis based on the FEMA P-58 methodology (FEMA, 2012). Ratings are established for the three categories of performance including safety, repair cost and functional recovery time. The results showed that both buildings achieve ratings of five and two stars for the safety and recovery dimensions respectively. However, for damage dimension, 2A gets a four stars rating while 2B has a five stars rating. At the design basis earthquake (DBE) level, the mean repair cost normalized by the building replacement value is 5.57% for the code-based building and 4.01% for the LATBSDC building. However, at the MCE level, the repair costs for the LATBSDC building (20%) is about 18% higher than that of the code-based building (17%). This result is explained by the fact that the residual drift demands are significantly higher in LATBSDC building and dominates the losses at the MCE level. For both buildings, the REDi recovery time is dominated by impeding factors which account for more than 75% of the functional recovery time.

Structures are generally under the risk of earthquake events, especially in regions with high seismicity. However, rather than occur individually, the seismic events are tend to happen in sequence of the mainshock and the multiple aftershocks, which would cause additional damage to the damaged buildings in post-mainshock environment and increase the probability of collapse. This study mainly focuses on characterizing the seismic risk of a 4-story steel frame under mainshock-aftershock earthquake sequences. The damage of the structure are captured using the maximum story drift ratio and the energy-based damage indices, of which the performances are evaluated and compared in prior. The result shows that the energy-based damage index has higher ability to capture the potential structural damage and tend to display higher rate of collapse in the lifetime of buildings.

Quantifying and propagating aleatory and epistemic uncertainty in nonlinear structural response simulation is key to robust performance-based seismic assessments. In this thesis, the focus is on the probabilistic seismic performance assessment of a non-ductile three-story reinforced concrete infilled frame building. The uncertainty in ground motion records and the uncertainty embedded in structural model parameters are explicitly considered. An equivalent strut model is used for the masonry infill walls, where the six constitutive parameters that define its backbone curve are treated as random variables. The variability in these parameters is characterized by developing correlated and uncorrelated distributions of the deduced-to-predicted ratios using data from 113 experimental tests. The uncertainties are propagated using Latin hypercube sampling to generate randomized structural model realizations. Multiple stripe analysis is performed with hazard-consistent ground motions. The effect of considering modeling uncertainty is investigated in terms of the distributions of maximum story drift ratios and drift based fragility functions. It is shown that the inclusion of modeling uncertainty has significant effects on the seismic performance of the case study building. The dispersion in the response and the mean annual frequency of exceeding the drift-based limit states are increased when modeling uncertainty is included. The initial stiffness of the infill walls is observed to have the most significant contribution to the performance of the building.

Controlled rocking steel braced frames (CRSBFs) have been developed with the goal of minimizing the post-earthquake impact of primary building functions. While there has been significant research to date to demonstrate the viability of the CRSBF as a high-performance system, much less has been accomplished in the development of performance-based design and assessment methods. This research is focused on developing models, tools and techniques for practicing engineers to analyze, design and assess the performance of CRSBFs. To avoid the computational expense of nonlinear response history analyses, an approximate method is formulated to estimate the CRSBF drift demands using the primary design parameters. Additionally, a reliability-based methodology for establishing the load and resistance factors for the force-controlled (braced frame) members is formulated. A key departure from previously developed capacity design approaches is the development of an explicit link between the effect of the failure of the force-controlled components and system level performance limit states (collapse and post-earthquake structural safety). The results from a case study applied to 3-, 6- and 9-story building cases show that the effect of force-controlled components is more significant for the collapse limit state compared to post-earthquake structural safety. Also, even when the resistance to load factor ratio (ϕ/γ) is increased to 1.8, the 50-year collapse probability remained below the 1% threshold prescribed by current building codes. The effect of record-to-record and modeling uncertainty on the seismic response and performance assessment of CRSBFs is also studied. The results showed that the impact of modeling uncertainty on seismic performance increases with the building height. To enable practitioners to estimate the service life costs of potential designs, surrogate models are developed to assess earthquake-induced life cycle economic loss and environmental impacts. The effectiveness of the surrogate models is demonstrated by evaluating their accuracy on “unseen” (i.e., not used in the development of the surrogate models) designs. 

Housing plays a primary role in post-earthquake recovery because all sectors of the economy rely on residents having healthy living conditions so that they will remain in the affected region. Therefore, understanding the time-dependent effects of earthquake events on housing is critical for improving post-earthquake trends through policy and planning interventions. This study develops an integrated framework for modeling post-earthquake housing recovery that combines probabilistic building performance with the decisions, actions and socioeconomic vulnerability of the affected populations. Building performance is characterized using limit states such as post-earthquake occupiability and repairability, which are explicitly linked to community functionality and recovery. Fragility functions are used to link the probability of exceeding these limit states to ground shaking intensities. Household decisions are modeled using empirical probabilistic utility models. Probabilistic discrete state models are implemented to represent post-earthquake recovery trajectories at the building, neighborhood and community scales. These models, which are inherently stochastic, can be purely empirical, such that the temporal parameters are sampled from an appropriate probability distribution. Alternatively, a simulation-based model can be employed to explicitly account for the effect of resource-availability on the time to complete relevant recovery activities. Socioeconomic vulnerability and other exogenous (external to household) and endogenous (external to household) factors are statistically linked to the temporal recovery parameters. The proposed model can be used to assist policy-makers, municipal governments and planners in understanding the possible interdependencies, interventions, and tradeoffs associated with post-earthquake housing recovery.

The increase in seismic activity after a large-magnitude earthquake coupled with the reduction in the lateral load-carrying capacity of the affected structures presents a significant human and financial risk to communities. The focus of this study is placed on formulating a framework for quantifying the impact of both the elevated post-mainshock seismic hazard as well as the mainshock-induced structural damage on building seismic performance. The viability of the proposed framework is examined through its application to mainshock-aftershock seismic performance evaluation of a set of reinforced concrete frames. Additionally, discrepancies between the frequency content of mainshock and aftershock ground motions and their impact on the seismic performance of RC moment frames is investigated. Two metrics are used in evaluating the seismic performance; seismic risk and seismic-induced financial losses. Both metrics are evaluated in both pre- and post-mainshock environments. The time-dependent nature of seismic hazard in the post-mainshock environment is accounted for through the adoption of a Markov risk assessment framework. In the post-mainshock environment, the seismic risk is examined as a function of the time elapsed since the mainshock’s occurrence while in the pre-mainshock environment, the risk is investigated during an assumed lifespan of 50 years for the studied structures. For the buildings and the high-seismicity site used in this study, both the increased post-mainshock seismic hazard as well as the reduction in the structural capacity are found to have a great influence on the seismic risk. The application of the proposed frameworks for seismic performance under mainshock-aftershock ground motions to the reinforced concrete frame buildings in Los Angeles County is also demonstrated. The outcomes of the regional seismic performance analysis show that omitting aftershocks from the seismic performance steps would lead to underestimating annual expected seismic risk and loss by up to 50% and 15%.