Earthquake Engineering & Structural Dynamics最新文献

Machine-learning models play a crucial role in structural seismic risk assessment and facilitate decision-making by analyzing complex data patterns. However, the dynamic nature of real-world data introduces challenges, particularly data drift, which can significantly affect model performance. This adversely affects machine-learning models intended to aid emergency responders and disaster recovery teams. This study primarily focused on assessing the impact of column corrosion-induced data drift on the performance of machine-learning models for seismic risk assessment of bridges. The machine-learning model performance was evaluated with and without considering the impact of corrosion. The results revealed a significant decrease in prediction accuracy when the data drift effect was not considered. To address this challenge, this study proposes integrating principal component analysis-based anomaly detection to enhance the model performance. The optimized model considering drift demonstrates significant improvements in accuracy across corroded bridges aged 25, 50, and 75 years, with accuracy rates increasing from 90%, 85%, and 81% to 98%, 97%, and 96%, respectively.

The study presents experimental and numerical results on two-dimensional and three-dimensional full-scale exterior non-seismically designed (NSD) reinforced concrete (RC) beam-column joint subassemblies subjected to quasi-static cyclic lateral loading. The tests were augmented by detailed 3D finite element modeling to obtain further information about the joint behavior. Through these systematic investigations and their detailed evaluation, clear conclusions could be drawn on the effect of transverse beams and slab on the overall seismic behavior of beam-column joints, where the joint core was devoid of transverse reinforcement. It was found that the presence of transverse beams enhanced both the ultimate joint shear strength and joint shear strength at first joint cracking. The crack development in concrete revealed that the diagonal joint shear cracks extended from the joint core into the transverse beams. The slab participation under flexure, when acting in tension, decreased with increase in drift due to intervening loss in joint stiffness, which was inconsistent with the observations in subassemblies where the joints were confined with transverse reinforcement. It was found that the inclined cracking in the transverse beams was caused due to joint shear stresses and aggravated due to torsional stresses when a slab was present. Normalized joint shear stress and principal tensile stress values were evaluated for first joint shear cracking and ultimate joint shear strength. These values may be useful for the seismic assessment of non-seismically designed beam-column joints with transverse beams and slab.

Identifying near-fault pulse-like ground motions from extensive ground motion databases holds paramount importance, as it provides a pivotal foundation for further inquiries into this specific type of ground motions, including the modeling of such stochastic processes as well as thorough analysis of their potential impact on structures and infrastructure systems. Currently, a diverse array of quantitative methods for identifying pulse-like ground motions have emerged, all of which demonstrate good accuracy within their respective research scopes. However, due to the limitations of each individual method in identifying specific cases, these diverse approaches often yield inconsistent results for certain ground motion records, posing a significant challenge in establishing a reliable classification criterion that relies solely on a single identification method. To address this issue, the present study adopts a multifaceted approach. Instead of improving a single time-frequency analysis-based identification method, it carefully conducts a selection of seven baseline methods through a systematic overview of the field. By leveraging the analytic hierarchy process (AHP), a comprehensive categorization method is developed that integrates the strengths of each approach, resulting in a more robust and credible classification criterion. According to the devised category indicator, ground motions can be classified into four categories: Category A comprises definitively pulse-like ground motions; Category B comprises apparently pulse-like ground motions; Category C consists of probably pulse-like ground motions; and Category D encompasses ground motions unlikely to exhibit pulse-like characteristics. It provides a more elaborate classification beyond the binary distinction of pulse-like and non-pulse-like ground motions associated with traditional onefold classification methods. For validation purposes, a basic dataset comprising near-fault ground motion records from the NGA-West 2 database has been utilized. To verify the comprehensive categorization method, two datasets of pulse-like ground motion records suggested by FEMA and PEER and one dataset of ground motion records collected during the 1999 Chi-Chi earthquake are addressed. Numerical examples illustrate the remarkable effectiveness of the proposed method in identifying near-fault pulse-like ground motions based on their varying degrees of pulse-like characteristics.

The seismic risk mitigation plans are vital since vulnerable structures are prone to partial or total collapse under the effect of future major earthquake events. Therefore, vulnerable structures in large building stocks should be determined using robust and accurate methods to prevent loss of lives and property. In the current state-of-the-art, the risk states (i.e., whether risky or not) of structures completely depend on the experience of the reconnaissance team of engineers, which could not result in standardized decisions. In this study, machine learning has been integrated into the decision-making algorithm to classify more precise and reliable seismic risk states of masonry buildings, categorizing them into up to four risk categories. For this purpose, a large database, including 12 features and detailed seismic risk analysis results of 4356 masonry buildings, is formed. Firstly, the input variables are preprocessed using feature engineering methods. Then, several machine learning algorithms are utilized to produce a network to estimate the risk state of masonry buildings in association with the risk states obtained from the detailed analysis results. As a result of the analysis of these algorithms, the correct prediction percentages for the testing database of the proposed method for two, three, and four risk states classification are predicted as approximately 87.5%, 86.6%, and 79.0%, respectively. This new approach makes it possible to produce risk color maps of large building stocks and reduce the number of buildings that require immediate action.

Ultra-high voltage (UHV) converter stations are critical nodes in power grids. This paper proposes a probabilistic framework for assessing and mitigating the seismic risk of UHV converter station systems to enhance the seismic performance of the grid. First, a Bayesian network model for the system functionality of UHV converter stations was established based on the enumeration of equipment failure scenarios. Conditional probability tables (CPTs) were used to represent the causal relationship among subsystems and system functionality. Inference calculations were conducted using Bayes’ theorem. Then, the definition of system seismic loss risk distribution was proposed to assess the seismic risk of the system over its entire lifespan. The feasibility of this framework was validated using a specific UHV converter station, yielding analytical solutions for the probability distribution of system functionality and seismic vulnerability curves. Additionally, the cost-effectiveness of several risk mitigation strategies was assessed. A cost-benefit analysis was performed from the perspectives of both the expected loss of a single earthquake and the life-cycle cost. The framework comprehensively considered the constraints imposed by series, parallel, and bypass control devices on the system's functionality. It was revealed that the seismic loss risk for UHV converter stations exhibits a characteristic of low probability but high loss.

This paper introduces a novel method to spontaneously predict displacement time histories and hysteresis curves of bridge lead rubber bearings under seismic loads and axial forces. The method leverages a stacked convolutional-bidirectional Cuda Long Short Term Memory network, enhanced with multi-head attention, skip connections, exponential learning rate scheduler, and a hybrid activation function to improve performance. The framework utilizes the functional application programming interface provided by the Python Keras library to build a model that takes input features such as horizontal and vertical ground accelerations, actuator loads in both lateral and vertical directions, and the superstructure mass. The effectiveness of the deep learning model is evaluated using a considerable experimental dataset of 53 real-time hybrid simulations, spanning various earthquake intensities and superstructure masses (Chi-Chi: 15 scenarios, El Centro: 15 scenarios, Kobe: 13 scenarios, and Northridge: 10 scenarios). Initially, Northridge earthquake data serves as unseen data, while the rest is used for training and validation. In a subsequent trial, the unseen data is centered on Kobe earthquake scenarios. By employing a hybrid loss function merging mean square and mean absolute errors, the model exhibits a substantial correlation of over 83% between predicted displacement time series and empirical measurements for the unseen data. In summary, the proposed model offers miscellaneous benefits, including time and cost savings in experimental efforts by decreasing the need for additional tests. It further delivers a swift and precise insight into the bridge bearing performance and its energy dissipation, facilitating timely and accurate bridge design in different scenarios for engineers.

This paper presents an experimental study on an innovative timber-based retrofit solution for reinforced concrete (RC) framed buildings, with or without masonry infills. The intervention aims to enhance seismic resistance through a light, cost-effective, sustainable, and reversible approach integrating energy efficiency upgrades. The method employs cross-laminated timber (CLT) panels as infills or external retrofitting elements, mechanically connected to the RC frame through steel fasteners. The system is combined with thermal insulation for improved energy efficiency. The seismic performance of the proposed retrofit technique was assessed experimentally on a full-scale building model at the European Laboratory for Structural Assessment (ELSA). The experiments included tests on two five-story building configurations: a masonry-infilled RC building as a reference and the same structure strengthened with CLT panels. Each building was subjected to unidirectional earthquake simulations of increasing intensity using the pseudodynamic (PsD) testing method with substructuring. The physical substructure of the hybrid model consisted of the first story of a two-story mockup built and retrofitted in the laboratory, while stories two to five were simulated numerically. The paper discusses major observations from the tests, comparing the damage evolution and hysteretic responses of the two configurations. The experiments yielded promising results, showing that the suggested retrofit solution significantly increased displacement and energy dissipation capacity. The retrofitted building survived earthquake intensities up to 50% higher than the non-retrofitted counterpart, exhibiting only slight structural damage. These pioneering experiments provide compelling data on the high effectiveness of the proposed CLT-based retrofit system in enhancing the seismic performance of full-scale RC buildings.

Real-time hybrid simulation (RTHS) technique significantly streamlines experimental procedures by allowing researchers to study a substantial portion of the structure through numerical analysis. For effective real-time interconnectivity between the investigated substructures, the numerical component must be solved within an extremely tight time frame. However, achieving a real-time solution for large numerical substructures presents a major challenge. Hence, this paper proposes the Proper Orthogonal Decomposition (POD) method to reduce computational burden in RTHS and shows its implementation. The merits of the approach are shown by comparisons between the full-order and reduced-order numerical substructures, including nonlinearities. A shear frame retrofitted with superelastic shape memory alloy dampers is investigated as a numerical model. The soil-structure interaction is also included using a finite element half-space model with an artificial viscous-spring boundary. Furthermore, the numerical substructure is coupled with shaking table experiments of a tuned liquid column damper to prove the feasibility of the method. With POD, the studied nonlinear numerical substructure can simulate up to 2655 degrees-of-freedom (DOFs) with a given hardware setup, while the full-order model is limited to 135 DOF, underscoring the significance of the POD method in RTHS.

Seismic fragility analysis and resilience assessment of large-span cable-stayed bridge structures are critical for evaluating their seismic performance. However, there is a scarcity of research on the effects of multi-support ground motions and their non-Gaussian characteristics on seismic fragility and resilience. This paper aims to addresses this issue. Initially, random ground motions with spatial variability and non-Gaussian characteristics are simulated using the Spectral Representation Method (SRM) and the Unified Hermite Polynomial Model (UHPM). Subsequently, the Fractional Exponential Moments-based Maximum Entropy Method (FEM-MEM) and the Adaptive Gaussian Mixture Model (AGMM) are employed for seismic reliability-based fragility analysis, overcoming the shortcomings of conventional lognormal assumption. Component- and system-level fragility analyses are conducted sequentially, followed by seismic resilience assessment of bridge structures based on the results of system-level fragility analysis. A numerical example is presented to validate the proposed method. Computational results indicate that: (1) The proposed method offers higher accuracy and broader applicability for seismic fragility analysis of large-span cable-stayed bridge structures compared to traditional assumptions. (2) The non-Gaussian characteristics of ground motions may significantly impact the seismic fragility analysis and resilience assessment of large-span bridge structures.

Substations function as neural hubs within power systems and play pivotal roles in the aggregation, transformation, and distribution of electrical energy. Previous experiences indicate that substation systems are highly susceptible to damage under earthquakes, resulting in a subsequent decrease in power supply functionality. To mitigate the risk of earthquake-induced damage, a novel approach based on Bayesian theory is proposed to assess the seismic vulnerability of complex engineering systems. The proposed method initially obtains the prior distribution of seismic fragility parameters for electrical equipment through numerical simulations of coupled finite element models. Subsequently, seismic damage survey data and Bayesian updating rules are applied to update the prior probability, obtaining a hybrid fragility function for electrical equipment. The Bayesian network was constructed using logical relations among internal electrical components in the substation, aiming to quantify the seismic vulnerability of the system across different functionality indicators. Finally, the causal inference technique was employed to quantify the importance of various components and equipment. A realistic case study on a typical 220/110/35 kV substation system was performed using the proposed method. The results demonstrate that the method improves the confidence level of the equipment fragility curves, reduces the computational workload of the system vulnerability analysis, and provides a theoretical basis for improving substation performance and formulating post-disaster maintenance plans.