Prediction Of Seismic Response Research Articles

Appropriate intensity measures for probabilistic seismic demand estimation of steel diagrid systems

Performance-based earthquake engineering (PBEE) predicts the performance of structures subjected to earthquake ground motions to improve seismic risk mitigation decisions. In PBEE, intensity measures (IMs) are employed as a link between earthquake hazard and seismic response. To the best of the authors' knowledge, investigation of the appropriate IMs for the probabilistic seismic demand evaluation of steel diagrid systems has not been presented in literature so far. In this study, the efficiency, practicality, proficiency, sufficiency, and scaling robustness of 38 candidate IMs for estimating the seismic response of low- to high-rise steel diagrid systems are investigated. Finite-element models of the structures are performed in OpenSees software and subjected to incremental dynamic analysis (IDA) using 22 far-field earthquake records. The whole response of the investigated diagrid systems is captured in terms of maximum inter-story drift ratio (MIDR). The appropriate IMs among the candidate IMs for steel diagrid systems are identified utilizing the concepts of efficiency, practicality, proficiency, sufficiency, and scaling robustness. The regression analysis results showed that peak ground velocity (PGV) and the ratio of PGV to peak ground acceleration (i.e., PGV/PGA) are the most appropriate IMs for seismic response prediction of steel diagrid systems under far-field ground motions. Furthermore, the collapse fragility curves, as well as drift hazard curves, were obtained based on various IMs. It was observed that the proposed IMs reduce the uncertainty in fragility curves. Also, it was concluded that seismic IM has a significant impact on the structural response hazard of steel diagrid systems.

Seismic Performance Evaluation of Two-story and Three-span Subway Station in Different Engineering Sites

ABSTRACT This study firstly investigates the seismic intensity measures (IMs) from the existing scalar ones that are suitable for the prediction of seismic response of shallowly buried rectangular underground structures. The peak acceleration and peak velocity at the ground surface are identified as the optimal IMs based on their efficiency, practicality, and proficiency. A series of seismic fragility curves are then developed for a two-story and three-span underground subway station embedded in three different typical engineering sites using nonlinear incremental dynamic analyses. In this study, the input ground motions at the level of engineering bedrock for the underground structure–soil interaction system are obtained through one-dimensional equivalent linear site response analysis according to the one-dimensional wave propagation theory. Comparison of the numerical results in this paper with the previous empirical and numerical fragility curves shows that the fragility analysis method in this paper is feasible, which can quantitatively give the failure probability of structures at different performance levels, and can provide reference for seismic design of underground structures. The numerical results indicate that both the characteristics of ground motions and the site profile have significant influence on the seismic fragility curves of underground structures. Underground subway stations are generally more vulnerable to seismic damage when embedded in engineering sites with lower average shear velocity. Moreover, seismic damage margin ratios are proposed for the underground structure in different site classes to show the confidence levels of its seismic performance when subjected to different levels of earthquake events, and can be used as a preliminary index for the earthquake risk assessment of subway station.

SIMPLIFIED ANALYSIS AND SEISMIC RESPONSE OF INTER-STORY ISOLATION SYSTEMS WITH AN ATTACHED INERTER

The inter-story isolation technology is a new control method in the field of earthquake engineering. The analysis model of the inter-story isolation hybrid control system with an attached inerter is established. The dynamic characteristic formulae of the inter-story isolation hybrid control system with an attached inerter are derived. The influence of the calculated parameters of the model on the dynamic characteristics is analyzed. The seismic response prediction formula based on the response spectrum is established and verified by time-history response analysis. The response control effect of the hybrid control system with an additional inerter and the influence of key parameters on control effect are analyzed. The results show that the inerter has a certain contribution to the control of the seismic response of the inter-story isolation structures. The proposed method can be used to predict the seismic response of the inter-story isolation hybrid control system with an attached inerter. The inerter can realize the dual control target of both displacement and acceleration. The results provide a reference for the simplified analysis methods and hybrid control of inter-story isolation systems.

Investigating seismic behavior of horizontally curved RC bridges with different types of irregularity in comparison with equivalent straight bridges

In this paper, the seismic behavior of a total of 84 reinforced concrete bridges, including 12 straight and 72 curved bridges was studied. The bridges were designed and detailed according to the American Association of State Highway and Transportation Officials (AASHTO) regulations. To evaluate the accuracy of the seismic response predictions obtained using elastic analyses, as suggested by AASHTO, both linear and nonlinear time history analyses were performed and the results were compared. In addition, to investigate the accuracy of the code provisions regarding the limitations on the bridge subtended angle for the use of equivalent straight bridges, the results obtained for the curved bridges were compared with those obtained for the equivalent straight bridges. Furthermore, the effects of different types of irregularity on the predicted displacement demands from equivalent straight bridges were investigated. To study the effects of different types of irregularity on the results, different irregular arrangements of spans, various column heights, different abutment conditions, and different subtended angles were considered. Also the effect of combination of different types of irregularity was investigated by considering bridges with different configurations. The results indicated that the deviation of the results obtained from the linear and nonlinear analyses is significant, particularly in the transverse direction of the bridges with irregularity in columns height for different subtended angles. Also, it was shown that in bridges with restrained abutments, unlike the bridges with unrestrained (i.e., free) abutments, the equivalent straight bridges cannot be used in lieu of the curved bridges, since the differences between the results from the two models are significant and become larger with increasing the subtended angle of the bridges. It was shown that some modeling assumptions widely used for seismic analysis of bridge structures can be unrealistic and may lead to inaccurate predictions. It was shown that the AASHTO specifications regarding the use of elastic analysis for some irregular bridges and the limitations on the subtended angle for the use of equivalent straight bridges need to be reevaluated.

Nonlinear lateral displacement of the single pile – A seismic analytical investigation

In the construction industry, pile uses to enhance seismic stability of infrastructure and tower, if the soil faces insufficient bearing capacity and unallowable nonlinear lateral and vertical displacements. In this study, I assume that the nonlinear lateral displacement mechanism and its prediction were analyzed and developed based on the non-linear seismic load. (i) The nature of seismic load, (ii) length of pile, and (iii) soil elastic modulus become model parameters to assess the nonlinear lateral displacement and vibration mechanism of a single pile. In analytical processes, instead of using the whole seismic load for calculation nonlinear lateral displacement of the pile, a method was suggested for minimization of the analytical process. The statistical models and analysis have been applied for evaluating the accuracy of the analytical results and monitoring distribution probability of nonlinear lateral displacement of pile concerning different factors. The results illustrate the effect of the directions of seismic load, more specifically their combination, as it provides the non-linearity nature of seismic load, length of pile, and soil elastic modulus govern nonlinear lateral displacement mechanism and, these factors are important for predicting nonlinear displacement and monitoring mechanism vibration of the pile. The suggested method minimizes the analytical process, provides acceptable single pile seismic response prediction, and enhances the accuracy of the analytical results. The presented method can apply to solve many nonlinear dynamics problems, and it is not limited to single pile seismic analysis and design.

Prediction of seismic drift responses of planar steel moment frames using artificial neural network and extreme gradient boosting

This study aims to develop machine learning (ML) models that can predict the seismic responses of planar steel moment-resisting frames subjected to ground motions. For this purpose, two of the most powerful ML techniques, artificial neural network (ANN) and extreme gradient boosting (XGBoost), were applied. To generate a comprehensive dataset for the training and testing of the ML models, 22,464 nonlinear dynamic analyses were conducted on 36 steel moment frames with different structural characteristics (i.e., number of stories, number of bays, column-to-beam moment capacity ratio) subjected to 624 ground motions. The selected ground motions had peak accelerations greater than 0.2 g and earthquake magnitudes of at least 4.0. The maximum top and interstory drifts were considered as the primary seismic responses of the steel frames. The ANN and XGBoost models were developed using MATLAB and XGBoost 1.1.1 Package, respectively. The results suggest that both the ANN and XGBoost models could reliably estimate the seismic drift responses of the steel frames. The XGBoost model achieved a better prediction than the ANN model in all the considered cases. The coefficients of determination (R2) of the XGBoost model for the testing dataset are 0.975 and 0.962 in the maximum top and interstory drifts, respectively. The significance of the input variables on the prediction of the seismic drift responses was also analyzed. The ground motion intensities (e.g., peak ground acceleration, velocity, displacement) had a more significant impact on the seismic drift response prediction compared with the earthquake (e.g., magnitude) and soil characteristics, and the spectral accelerations in the first three natural periods. In particular, the peak ground velocity had the most significant effect on the seismic drift responses prediction. Finally, a graphical user interface based on the XGBoost model was developed for the preliminary estimation of the seismic drift responses of steel moment frames.

Deep learning techniques for predicting nonlinear multi-component seismic responses of structural buildings

This paper presents a new approach for nonlinear multi-component seismic response prediction of structures using hybrid deep learning techniques. Modern recurrent neural networks map the relationship between acceleration time-series of the base/ground of a building and the superstructure, as a form of nonlinear time-history analysis method. Seismic responses were measured in three components which enables multi-component seismic predictions with adequate deep learning architectures. While long short-term memory (LSTM) neural networks can obtain data from a single component per surrogate model, hybrid convolutional-LSTMs (ConvLSTM) neural networks are utilized for multi-component purposes. A guide for pre-processing data and structuring the architecture of deep neural networks are proposed. Also, two filtering methods are compared, Fast Fourier Transform (FFT) Butterworth filter and discrete wavelet transform (DWT) decomposition. Decimation is implemented to reduce the features to useful values, as a dimension reduction approach. With enhancements to the architecture of the network, training time can be reduced significantly, and accuracy could be further improved. A challenging case study is addressed that covers an industrial level practical building. Results show that proposed hybrid models can even predict the capacity curves of a structure indirectly, providing new prospects for engineers to evaluate the seismic performance of a building.

Recurrent neural networks for complicated seismic dynamic response prediction of a slope system

Earthquake-induced landslides have resulted in huge casualties and considerable financial repercussions, and slope dynamic response analysis has always been a hot issue. The prediction of the dynamic response of a slope before the occurrence of future earthquakes will benefit disaster prevention and reduction. Traditional methods (such as the finite element method) are mostly based on simplified physical mechanisms and cannot accurately predict the dynamic response of complicated slope systems. This article innovatively applies novel recurrent neural networks to the prediction of the slope dynamic response. Using the results of large-scale shaking-table tests, we introduced a moving-steps strategy and established three recurrent neural network models: Simple-RNN, LSTM and GRU models. Moreover, a multi-layer perceptron prediction model was trained for comparative verification. We also conducted three experiments to investigate the effect of the data volume on the models. Results show that recurrent neural networks perform well in the analysis of the seismic dynamic response of a slope and provide better predictions than the multi-layer perceptron network. When there are many data, the LSTM and GRU models have advantages, and the confidence indexes of their predictions with normalized error within ±5% are 84.5% and 86.4%, respectively. It is concluded that recurrent neural networks are suitable for the time-series prediction of dynamic responses to seismic loads. To some extent, this paper may help reduce the future risks and losses of earthquake-triggered landslide disasters.

Multi-component deconvolution interferometry for data-driven prediction of seismic structural response

Prediction of structural response is necessary for evaluating condition and quantifying vulnerability of structural systems exposed to seismic loads. Traditional modeling techniques for infrastructure systems such as finite elements are typically limited by inherent modeling assumptions as well as the prohibitive computational effort required for analysis. This necessitates the development of surrogate models that serve as a basis for predicting structural response. Deconvolution interferometry is a viable data-driven approach for such a task that uses single component sensor data to generate a set of impulse response functions for a structure of interest that constitutes the required surrogate model of the structure. The resulting surrogate model aids in both dynamic characterization as well as for accurate response prediction. However, it is limited to cases where motions in various degrees of freedom of a structure can be decoupled. This decoupling requires dense sensor deployment as well as prior knowledge about the structure’s geometry. To overcome this limitation, in this paper we propose a multi-component deconvolution seismic interferometry approach to develop a surrogate model for response prediction for cases with sparse sensor deployment and limited information about the structure of interest. The resulting model incorporates various sources of uncertainties namely measurement noise, effects of variations of temperature and humidity, and human activity induced vibrations by predicting a probabilistic structural response. We demonstrate the efficacy of the proposed algorithm by applying it to field monitoring data collected from structures with sparse sensor deployment in the Groningen region of the Netherlands for a period of approximately 10 months on average.

High-fidelity broadband prediction of regional seismic response: a hybrid coupling of physics-based synthetic simulation and empirical Green functions

The prediction of the seismic response of critical structures is highly sensitive to many aspects, among which the earthquake source and the geological setting are prominent. The related uncertainty issues must be taken into account in seismic risk mitigation studies, for example through the evaluation of several realizations of a future earthquake scenario. This aspect is crucial when addressing vulnerability studies at a regional scale. When opting for physical-based numerical simulations (PBS), however, the computational burden increases along with the expected degree of fidelity, making it difficult to evaluate more than a few dozens of alternatives. To cope with this disadvantage, in this work an alternative method is proposed, which exploits a rather low number of synthetic earthquake simulations and combines them with the empirical Green function (EGF) method, to finally generate thousands of alternative yet realistic seismic response of the site of interest. This hybrid strong motion predictions benefit of both (i) PBS high fidelity and (ii) data assimilation of strong ground motion records in the seismic area of interest, via EGF method, producing broadband synthetics at a relative cheap computational price. The power of the hybrid method is tested on a real case scenario, embodied by the ground-shaking prediction at the nuclear site of Cadarache, in the surroundings of the fault of Middle Durance, in South-Eastern France. Thousands of broadband (0–15 Hz) hybrid synthetic seismic response are generated, associated with different fault parameters (EGF method) and based upon a few key physics-based simulations, accurate up to 5 Hz.

Seismic response prediction and variable importance analysis of extended pile-shaft-supported bridges against lateral spreading: Exploring optimized machine learning models

Reliable prediction of seismic responses of bridges against lateral spreading is critical for fragility and resilience assessment of transportation infrastructure. This problem, however, has remained a significant challenge due to the high complexity of the liquefaction phenomenon and its significance for the seismic performance of structures. The present study explores machine learning (ML) approaches, particularly for bridges supported by extended pile-shafts, for reliable estimation of bearing deformation and column drift ratio responses of bridges. The study considers a large set of covariates across soil, structural, and ground motion features. Five ML algorithms are examined including the traditional multiple linear regression (MLR) as the baseline reference, as well as Lasso regression, neural network (NN), random forests (RF), and gradient tree boosting (GTB). A large number of nonlinear soil-bridge finite element models considering the soil and structural uncertainties are dynamically analyzed under 720 ground motions, and the optimal parameters of the ML models are determined via five-fold cross validation. The results indicate that NN and GTB can well predict the seismic responses of the studied soil-bridge systems, followed by RF, while Lasso and MLR are generally not able to yield reliable estimates. Furthermore, a variable importance analysis is conducted using the produced regression models. Results indicate that intensity measures (IMs) (particularly the spectral acceleration at 2.0 s) are generally more significant than the soil- and structure-related variables. In addition to the IMs, variables associated with the nonliquefiable crust layer (i.e., thickness, strength, and sloping angle) and the concrete strength and longitudinal reinforcement ratio are generally significant variables, whereas the column diameter and rebar yielding strength are relatively insignificant.

Seismic Performance Evaluation of a Roof Structure of a Historic Chinese Timber Frame Building

ABSTRACT This paper presents a finite element modeling method for a prediction of lateral resistance and seismic response of the roof structure of a historic Chinese timber frame building. Force-displacement relationships of straight tenon joints, Mantou tenon joints, and Dougong bracket sets are calibrated by experimental results. In-plane and out-of-plane resistance analyses have been carried out, and force-displacement relationships and stress contributions are obtained. Acceleration and displacement responses are attained and adopted for the seismic performance evaluation of the historic roof structure. Results show that the overall in-plane and out-of-plane hysteresis performances of the roof are poor and energy dissipation abilities are weak. The seismic responses of measuring nodes are close under the same seismic excitation. As the seismic excitation increases, peak responses of all measuring nodes increase. Dougong brackets and mortise-tenon connections dissipate large seismic energy under strong excitation, and seismic responses (acceleration and displacement responses) of the roof structure are significantly reduced. The maximum drift angles of the entire roof truss are found to approach 1/2500 rad, demonstrating that the roof structure has good integrity.

A multi-scale attention neural network for sensor location selection and nonlinear structural seismic response prediction

Monitoring and predicting seismic responses of a civil structure can be used to assess its behavior under dynamic loading and to determine its structural health condition. In practice, the employed number of sensors is generally limited by the cost and functionality issues. This paper develops a practical solution of four-stage procedures, focusing on prediction of seismic displacement responses at all building floors using acceleration measurements at the optimized sensor locations. In this paper, a novel multi-scale attention-based recurrent neural network is proposed. In particular, the attention mechanisms in the network effectively focuses on more relevant input data among bidirectional ground accelerations and multivariate acceleration responses. The seismic response data for training the developed neural network is generated by performing nonlinear time historical analysis of a three-dimensional finite element model. A floor displacement warping loss is designed to numerically measure the discrepancy between the prediction and the ground truth. A case study is performed using the numerical and real-world data of a high-rise building to systematically evaluate the prediction performance of the proposed methodology. Results demonstrate that the proposed method outperforms the compared state-of-the-art methods in terms of prediction accuracy and reliability.

SEISMIC INTENSITY MEASURES FOR THE DAMAGE EVALUATION OF CIRCULAR TUNNELS

The seismic intensity measure (IM) is a key parameter affecting the accuracy of seismic risk assessment, which plays an important role in the performance-based earthquake engineering framework. A reasonable IM can make the prediction of structural seismic responses more accurate. To study on seismic IMs which are suitable for the damage evaluation of circular tunnels, a two-dimensional finite element model was established for the nonlinear dynamic time history analyses of circular tunnel-surrounding rock interaction. The near-field ground motions without velocity pulses and far-field ground motions recommended by FEMA-P695 were used as the input in the numerical model. The overall lining damage indices in compression and in tension were used as the engineering damage measures to estimate the damage states of the lining. The applicability of 20 ground motion IMs commonly used in the engineering practice were investigated. The results show that based on the criteria of efficiency, practicality, proficiency and sufficiency, the best seismic IM for the damage evaluation of circular tunnels is the sustained maximum acceleration (SMA), and the second best is the acceleration spectrum intensity (ASI).

Investigating the Use of Natural and Artificial Records for Prediction of Seismic Response of Regular and Irregular RC Bridges Considering Displacement Directions

The seismic responses of continuous multi-span reinforced concrete (RC) bridges were predicted using inelastic time history analyses (ITHA) and incremental dynamic analysis (IDA). Some important issues in ITHA were studied in this research, including: the effects of using artificial and natural records on predictions of the mean seismic demands, effects of displacement directions on predictions of the mean seismic response, the use of 2D analysis with combination rules for prediction of the response obtained using 3D analysis, and prediction of the maximum radial displacement demands compared to the displacements obtained along the principal axes of the bridges. In addition, IDA was conducted and predictions were obtained at different damage states. These issues were investigated for the case of regular and irregular bridges using three different sets of natural and artificial records. The results indicated that the use of natural and artificial records typically resulted in similar predictions for the cases studied. The effect of displacement direction was important in predicting the mean seismic response. It was shown that 2D analyses with the combination rules resulted in good predictions of the radial displacement demands obtained from 3D analyses. The use of artificial records in IDA resulted in good prediction of the median collapse capacity.

An H/V geostatistical approach for building pseudo-3D Vs models to account for spatial variability in ground response analyses Part II: Application to 1D analyses at two downhole array sites

Common procedures used to account for spatial variability of shear wave velocity (Vs) in one-dimensional (1D) ground response analyses (GRAs), such as stochastic randomization of Vs or increasing small-strain damping, have been shown to improve seismic site response predictions relative to 1D GRAs where no attempts are made to account for spatial variability. However, even after attempting to account for spatial variability using common procedures, 1D GRAs often still yield results that are different than ground motions recorded at many downhole array sites. When 1D predictions differ from observations, the site is typically considered to be too spatially variable to effectively use 1D GRAs. While there is no doubt that some sites are indeed too variable for 1D GRAs, it is also possible that simple 1D analyses could still be effectively used at many sites if spatial variability is accounted for through a more rational, site-specific approach. In this study, an H/V geostatistical approach for building pseudo-3D Vs models is implemented to account for spatial variability in 1D GRAs. The geostatistical approach is used to generate a uniform grid of Vs profiles that have been scaled to match fundamental site frequency estimates from horizontal-to-vertical spectral ratio (H/V) noise measurements. In this article, 1D GRAs are performed for each grid point and the results are statistically combined to reflect the average site response and its variability. This 1D application is demonstrated at the Treasure Island and Delaney Park Downhole Array sites, where it is shown to produce superior fits to the small-strain recorded site response relative to existing approaches used to account for spatial variability in 1D GRAs. Using the proposed approach, we also investigate the lateral area that is likely influencing site response at each site and show that it could extend to significant distances (as much as 1 km) from the boreholes.

Development of Artificial Neural Network Model for Prediction of Seismic Response of Building with Soil-structure Interaction

Development of Artificial Neural Network Model for Prediction of Seismic Response of Building with Soil-structure Interaction

Determination seismic effect of buildings and structures, taking into account changes model of external relations within the life cycle

This research paper discusses the problem of the strain state formation in the bearing system through its lifespan. It is demonstrated that in the general case the seismic events (being quite rare natural acts) can impact the structures which, by the time an earthquake hits, will experience strains developed during the period preceding the earthquake. Consequently, the accurate seismic load calculation for the bearing system in view of the seismic events shall account for the strain state developed in the structures prior to the seismic event. Analysis of the customary computational techniques used in seismic load calculations for bearing systems and employing the response spectrum method reveals, that in the context of a single-stage calculation method it is possible to introduce only one model of external relations that corresponds to either the long-term loading effect or to the short-term seismic loads. Such approach is not aligned with the true-life pattern of external relations, which is subject to changes through the life span. For the purposes of accurate prediction of seismic response in buildings, it is proposed to use a multistage calculation method with the stage-to-stage inheritance of the stress-strain state. At that, at any stage it is possible to modify the properties of external relations. Compared are the calculation results of the seismic impact test model employing the customary calculation techniques and the multistage calculation method with adjustment of the external relations model to correspond with the lifespan mode. The outcome of the research proves that the use of the customary single-stage calculation method leads to significant simplification of the bearing system strain model, as well as to inaccuracy of the seismic response prediction result. Based on the research outcome, the single-stage calculation method is advised. It permits to take into account developments in the external relations model through the building lifespan.

Seismic response of large-span spatial structures under multi-support and multidimensional excitations including rotational components

To achieve rational and precise seismic response predictions of large span spatial structures (LSSSs), the inherent non-uniformity and multidimensionality characteristics of earthquake ground motions should be properly taken into consideration. However, due to the limitations of available earthquake stations to record seismic rotational components, the effects of rocking and torsional earthquake components are commonly neglected in the seismic analyses of LSSSs. In this study, a newly developed method to extract the rocking and torsion components at any point along the area of a deployed dense array from the translational earthquake recordings is applied to obtain the rotational seismic inputs for a LSSS. The numerical model of an actual LSSS, the Dalian International Conference Center (DICC), is developed to study the influences of multi-support and multidimensional excitations on the seismic responses of LSSSs. The numerical results reveal that the non-uniformity and multidimensionality of ground motion input can considerably affect the dynamic response of the DICC. The specific degree of influence on the overall and local structural displacements, deformations and forces are comprehensively investigated and discussed.

Collapse prognosis of a long‐span cable‐stayed bridge based on shake table test and nonlinear model updating

SummaryThe collapse of long‐span cable‐stayed bridges under strong earthquakes will not only result in severe casualties and loss of property but also significantly delay the rehabilitation of the affected area. It is therefore essential to study the failure progress and potential collapse mechanisms of these bridges under strong earthquakes for assisting their inspection and maintenance and helping them survive an earthquake. Collapse simulations in existing studies are commonly conducted with design‐document‐based finite element (FE) models, which usually neglect the actual damage state of the bridges. The influences of modelling uncertainties on the simulated responses cannot be addressed. This study proposes a nonlinear FE model updating‐based collapse prognosis method for long‐span cable‐stayed bridges, which includes (1) a correlated‐updating‐parameter‐based nonlinear FE model updating algorithm, (2) a restarted time history analysis‐based seismic response prediction method and (3) an elemental deactivation‐based collapse simulation algorithm. The shake table test of a scaled Sutong cable‐stayed bridge in China is adopted to illustrate the proposed earthquake‐induced collapse prognosis method. A refined FE model of the bridge is first established and nonlinearly updated using the measurement data. The collapse prognosis of the bridge under a subsequent strong ground motion is then performed based on the updated model. The predicted structural responses and final failure mechanisms are compared with the measured responses and experimental observations with good agreement, indicating that the proposed method is feasible and accurate for evaluating the seismic performance and failure mechanisms of long‐span cable‐stayed bridges.