USArray, a pioneering project for the dense acquisition of earthquake data, provides a semi-uniform sampling of the seismic wavefield beneath its footprint and greatly advances the understanding of ...the structure and dynamics of Earth. Despite continuing efforts in improving the acquisition design, network irregularity still causes spatial sampling alias and incomplete, noisy data, which imposes major challenges in array-based data analysis and seismic imaging. Here we employ an iterative rank-reduction method to simultaneously reconstruct the missing traces and suppress noise, i.e., obtaining free USArray recordings as well as enhancing the existing data. This method exploits the spatial coherency of three-dimensional data and recovers the missing elements via the principal components of the incomplete data. We examine its merits using simulated and real teleseismic earthquake recordings. The reconstructed P wavefield enhances the spatial coherency and accuracy of tomographic travel time measurements, which demonstrates great potential to benefit seismic investigations based on array techniques.
•A deep long short-term memory (LSTM) network is developed for nonlinear structural response modeling.•Two input-output schemes (LSTM-s and LSTM-f) are presented.•The deep learning model is capable ...of modeling both elastic and inelastic response of buildings.•An unsupervised learning algorithm is used to cluster the seismic inputs for training enhancement.•The approach was successfully verified by both numerical and experimental examples.
This paper presents a comprehensive study on developing advanced deep learning approaches for nonlinear structural response modeling and prediction. Two schemes of the long short-term memory (LSTM) network are proposed for data-driven structural seismic response modeling. The proposed deep learning model, trained on available datasets, is capable of accurately predicting both elastic and inelastic response of building structures in a data-driven fashion as opposed to the classical physics-based nonlinear time history analysis using numerical methods. In addition, an unsupervised learning algorithm based on a proposed dynamic K-means clustering approach is established to cluster the seismic inputs in order to (1) generate the least but the most informative datasets for training the LSTM and (2) improve the prediction accuracy and robustness of the model trained with limited data. The performance of the proposed approach is successfully demonstrated through three proof-of-concept studies that include a nonlinear hysteretic system, a real-world building with field sensing data, and a steel moment resisting frame. The results show that the proposed LSTM network is a promising, reliable and computationally efficient approach for nonlinear structural response prediction, and offers significant potential in seismic fragility analysis of buildings for reliability assessment.
•Effects of single (IP → OOP) and mutual (IP ↔ OOP) interaction of unreinforced MIs.•Effects of removal of MIs when collapse is induced by a combination of IP and OOP actions.•GIS aided structural ...mapping so as to represent typical residential infilled buildings.•Simplified shear-type tridimensional scheme of infilled framed structures.•Extensive numerical investigation in order to explore effects of the different design parameters.
In the aftermath of earthquakes poor seismic response was evidenced by non-structural elements such as enclosure masonry infills (MIs), characterized by a combination of in-plane (IP) and out-of-plane (OOP) damage mechanisms. The present work is aimed at identifying the effects of this mutual interaction on the nonlinear seismic behaviour of reinforced concrete (RC) framed structures. To this end, an extensive parametric study is carried out considering a spatial one-bay multi-storey shear-type model with MIs constituted of two leaves of clay hollow bricks, which is assumed as equivalent to infilled RC framed buildings. An extensive GIS aided structural mapping of the town of Rende (Italy) is adopted in order to obtain benchmark models as close to real structures as possible, focusing on total height, in plan dimensions and maximum and minimum bay lengths of the buildings. The dependence of the results on the variation of the following design parameters is considered for the i-th cluster of buildings characterized by the same number of storeys: i.e. fundamental vibration period; behaviour factor of the bare structure; aspect ratio of MIs, defined as the ratio between the panel length and height. A five-element macro-model comprising four (diagonal) nonlinear beams and one (horizontal) central nonlinear truss for the prediction of the OOP and IP behaviour of MIs, respectively, is first implemented in a homemade computer code. The proposed algorithm addresses the issue of the nonlinear mutual interaction of MIs by modifying stiffness and strength in the OOP direction on the basis of simultaneous or prior IP damage and vice-versa. A lumped plasticity model describes the inelastic behaviour of the RC frame members. Finally, a review of the current Italian, European and American seismic code provisions is performed by means of comparison with results of nonlinear dynamic analyses. To this end, bidirectional spectrum-compatible accelerograms are considered at ultimate limit states.
•The predictive expressions of moment and curvature limit states for high-strength reinforced high-strength concrete (HSC-HSSB) rectangular columns were proposed.•The predictive expressions of drift ...displacement and shear force limit states for HSC-HSSB rectangular columns were proposed.•Considering the uncertainty of structures, the standard deviation of limit states corresponding to drift displacement and shear force were proposed, which can be applied in fragility analysis.
This study presents the derivation of closed-form component limit state expressions for high-strength concrete columns reinforced with high-strength steel bars (HSC-HSSB) from an extensive database through the moment–curvature analysis approach. Initially, sensitivity analysis was conducted to investigate the regularity of the parameters (section size, axial-load ratio, longitudinal reinforcement ratio, stirrup reinforcement ratio, concrete compressive strength, and yield strength of longitudinal reinforcement) on the curvature limit state. The approach being convenient, massive samples were obtained in a short time, therefore, 6^7(279936) levels were generated to derive the predictive expressions of curvature and moment limit states by using the logarithm linearity model. Subsequently, based on the cross-section expressions, the drift displacement and shear force limit states predictive expressions were derived and verified by experimental studies, and these limit states were predicted with high precision. Moreover, after considering the uncertainty of material and section size, the curvature and drift displacement of the HSC-HSSB column exhibited a lognormal distribution, and the useful suggested standard deviations (βc) in different limit states were proposed for practical application of fragility curves to assess the vulnerability of bridge columns to earthquakes.
•Seismic pounding between multistory structures is likely to introduce a torsional behavior.•The inter – story pounding problem has not been studied effectively yet.•Asymmetric pounding between an ...8-storey RC building and an adjacent shorter one is investigated.•The column that suffers the hit of the adjacent structure proved to be the key issue in the inter – story pounding problem.•In asymmetric pounding neglecting its effects leads to non-conservative building design or even to critical situation.
Collisions between structures due to seismic oscillations have been reported many times in literature as a common case of damage. Further it is quite usual seismically induced oscillations of a structure in a city center block of buildings to be partly restrained in lateral displacements and therefore torsional behavior to be introduced in the structure. Two different types of structural interaction may be defined: (a) Diaphragm-to-diaphragm collisions (Type A). (b) The floor levels of the two structures are different. Consequently during the seismic oscillations the diaphragms of the first one impact the columns of the other (Τype B or interstory pounding). In this work the cases of an 8-story reinforced concrete building that suffers pounding with an adjacent structure that has 1, 2, 3, 4, 5, 6, 7 or 8 stories are studied. Pounding occurs only in one (Case 1) or in two (Case 2) columns of the structures and since the other columns are free to move without restrictions a torsional behavior is introduced (asymmetric pounding). Moreover in Type B interaction these columns of the 8-story frame structure undergo impacts at a height equal to 2/3 of their deformable length from the diaphragms of the other structure. The influence of an initial distance between the two interacting structures on the torsion effect is investigated too. Nonlinear seismic step-by-step analyses are performed. More than two hundred pounding cases with torsional effect each one for three natural seismic excitations are studied. Results in terms of shear and ductility demands of the columns are presented and commented. Both types A and B yielded high torsional structural rotation. In interaction Type B it can be deduced from the cases under examination that the column that is endured the impact from the top floor of the other structure develops high shear demands that exceed the available capacity many times during the step-by-step seismic analysis. Moreover high ductility demands have been observed for this column. Finally it is concluded that for buildings that may undergo asymmetric pounding not taking it into account may lead under certain conditions to non-secure design or even critical situations.
Deep-Learning Inversion of Seismic Data Li, Shucai; Liu, Bin; Ren, Yuxiao ...
IEEE transactions on geoscience and remote sensing,
03/2020, Volume:
58, Issue:
3
Journal Article
Peer reviewed
Open access
We propose a new method to tackle the mapping challenge from time-series data to spatial image in the field of seismic exploration, i.e., reconstructing the velocity model directly from seismic data ...by deep neural networks (DNNs). The conventional way of addressing this ill-posed inversion problem is through iterative algorithms, which suffer from poor nonlinear mapping and strong nonuniqueness. Other attempts may either import human intervention errors or underuse seismic data. The challenge for DNNs mainly lies in the weak spatial correspondence, the uncertain reflection-reception relationship between seismic data and velocity model, as well as the time-varying property of seismic data. To tackle these challenges, we propose end-to-end seismic inversion networks (SeisInvNets) with novel components to make the best use of all seismic data. Specifically, we start with every seismic trace and enhance it with its neighborhood information, its observation setup, and the global context of its corresponding seismic profile. From the enhanced seismic traces, the spatially aligned feature maps can be learned and further concatenated to reconstruct a velocity model. In general, we let every seismic trace contribute to the reconstruction of the whole velocity model by finding spatial correspondence. The proposed SeisInvNet consistently produces improvements over the baselines and achieves promising performance on our synthesized and proposed SeisInv data set according to various evaluation metrics. The inversion results are more consistent with the target from the aspects of velocity values, subsurface structures, and geological interfaces. Moreover, the mechanism and the generalization of the proposed method are discussed and verified. Nevertheless, the generalization of deep-learning-based inversion methods on real data is still challenging and considering physics may be one potential solution.
Recent decades have seen increased interest in using the controlled rocking concept in seismic resisting systems. Unlike conventional systems, where lateral deformation of a member is achieved ...through the formation of plastic hinges in critical regions, in the rocking systems this is achieved through a gap opening mechanism. Due to gravity load and/or post-tensioning forces, the rocking systems exhibit a self-centering behavior. Conducting a continuum finite element analysis to investigate the seismic response of such a system is quite expensive in terms of computational resources. On the other hand, a simplified macro model using two springs to simulate the gap opening/closing mechanism cannot accurately predict the dynamic response of the system. This study utilizes a multiple-spring model to simulate the nonlinear seismic response of circular tubular steel piers. An efficient optimization procedure based on a genetic algorithm is developed to calibrate the parameters of the springs. The results of continuum finite element analyses are compared with those obtained from the multi-spring model to verify the accuracy of the model. The proposed method is shown to be advantageous for accurately simulating the seismic response of a bridge model subjected to multi-directional ground motions, particularly the hysteretic force-displacement relationship, and dynamic response time history.
•Seismic response of posttensioned (PT) rocking steel bridge piers is investigated through continuum and macro finite element (FE) modeling approaches.•Computationally efficient macro modeling approaches, i.e., two and multi -spring macro models, are discussed.•A procedure for calibrating the parameters of the multi-spring model using genetic algorithm is presented.•The performance of two-spring and multi-spring macro models in predicting the seismic response is examined.•Multi-spring macro model is extended to simulated the response of double rocking configuration.
•Modeled corrosion-fatigue degradation to assess life-long performance of RC bridges.•Rebar pitting corrosion and traffic-induced fatigue stresses form this degradation.•Degrading bridges are ...analyzed for earthquakes with substantial vertical components.•FE model of the bridge is updated at every life-cycle year to capture degradation.•Results portray higher life-cycle seismic vulnerability of all critical components.
Highway bridges, located in heavy traffic and corrosive environments, are prone to corrosion-fatigue degradation, due to which material properties of steel and concrete in bridge girders get altered and effective area of rebars reduces resulting in reduced flexural stiffness of girders. Consequently, exposed bridges become weaker to resist earthquake loads, particularly for the ones having three significant translational components. It is, therefore, important to evaluate life-long performance of bridges that suffer from corrosion-fatigue degradation and are subjected to seismic events having substantial vertical ground motions (VGMs). Such a threatening scenario for seismic safety of bridges has not been explored yet. The current study demonstrates the confronting role of corrosion-fatigue degradation and VGM on bridge seismic response through a representative multi-span RC bridge in Gujarat, India, and required information on corrosion, traffic, and seismic activities is acquired from appropriate sources. The degradation mechanism is modeled as a coupled phenomenon of pitting corrosion of rebar and traffic-induced fatigue stresses that give rise to fatigue cracks at deep pit locations of rebars in girders. Developed numerical model of the bridge composes of all essential features to accurately capture the gradual degradation along life-span and fluctuations in response of key vulnerable bridge components under horizontal and vertical excitations. Results suggest that ignoring the impact of this degradation mechanism and VGM can lead to higher seismic vulnerability of bridges and endanger their safety under future earthquakes.
The continuously growing amount of seismic data collected worldwide is outpacing our abilities for analysis, since to date, such datasets have been analyzed in a human-expert-intensive, supervised ...fashion. Moreover, analyses that are conducted can be strongly biased by the standard models employed by seismologists. In response to both of these challenges, we develop a new unsupervised machine learning framework for detecting and clustering seismic signals in continuous seismic records. Our approach combines a deep scattering network and a Gaussian mixture model to cluster seismic signal segments and detect novel structures. To illustrate the power of the framework, we analyze seismic data acquired during the June 2017 Nuugaatsiaq, Greenland landslide. We demonstrate the blind detection and recovery of the repeating precursory seismicity that was recorded before the main landslide rupture, which suggests that our approach could lead to more informative forecasting of the seismic activity in seismogenic areas.
PDF
