Sequential Ground Motions Research Articles

Fragility Analysis of Masonry Structures Subjected to Random Sequential Ground Motions

In this study, a fragility analysis of masonry structures is conducted using a random sequential ground motion model. Initially, the relationship between a mainshock and its aftershocks is clarified based on the physical mechanism of “source-path-local site”. Building on this, the random sequential ground motion model is developed by integrating a point-source model with a homogeneous isotropic medium model. A five-story masonry structure model is then constructed in ABAQUS (6.14), and its accuracy is validated through experimental testing. Following this, 21 sets of random sequential ground motions are generated. Using the Incremental Dynamic Analysis (IDA) method, 1260 nonlinear response samples of the structure under varying sequential ground motions are obtained, providing an insight into the fragility of the masonry structure subjected to these ground motions.

Using image-based inspection data to improve response predictions of earthquake-damaged unreinforced masonry buildings

AbstractExplicit representation of uncertainties is essential to improve the reliability of seismic assessments of earthquake-damaged buildings, particularly when dealing with unreinforced masonry buildings. Modern inspection techniques use images for detecting and quantifying the damage to a structure. Based on the principle of falsification, this paper evaluates how the use of information of damage that is obtained from images taken on earthquake-damaged buildings reduces the uncertainty when predicting the seismic response under a future earthquake. New model falsification criteria use information on the residual state of a building, such as shear cracks, residual roof displacements, and observation of out-of-plane failure. To demonstrate the effectiveness of these criteria in reducing the uncertainty in response predictions, results from a four-story unreinforced masonry building stiffened with reinforced concrete walls, which was experimentally tested under a sequence of ground motions, are assessed. Three commonly used modeling approaches (single degree of freedom (DOF) systems, multi DOF systems with four DOFs, and equivalent frame models) are used, where uncertainties in model parameters and model bias are included and propagated through the analysis. Out of the models used, and in the absence of any additional source of information, the proposed falsification criteria are most effective in connection with the equivalent frame model because this model can simulate the response at the element-level, while the simpler models can only represent the global response or the response at the storey-level. The results show that when using only the information on the presence of shear cracks, which might be the first and only source of information after an earthquake, the effectiveness of model falsification is increased, thus reducing the uncertainty in model parameter values and seismic response predictions through the use of image-based inspection.

Seismic fragility analysis of girder bridges under mainshock‐aftershock sequences based on input‐output hidden Markov model

AbstractCurrent seismic design codes for bridge structures do not account for the influence of aftershock sequences, which, to some extent, overestimate the seismic performance for bridges subjected to mainshock‐aftershock (MS‐AS) scenarios. To address the great need for ground motion sequences tailored to specific research sites for fragility analysis, this study proposes a method for generating artificial MS‐AS ground motion sequences based on the evolutional bimodal Kanai–Tajimi model and the Epidemic–Type Aftershock Sequence model. We establish a framework for MS‐AS fragility analysis using an input–output Hidden Markov Model (IOHMM), where the damage states (DS) of bridge piers are considered unobservable and are inferred statistically through damage indices in an unsupervised manner. Model parameters are trained using intensity measure (IM) sequences and damage index (DI) sequences. Fragility curves for both the mainshock and state‐dependent aftershocks considering multiple aftershocks are formulated based on the initial state probability and state transition probabilities of the proposed IOHMM. The fragility analysis results reveal that as the initial seismic damage level increases, the probability of aftershocks causing higher damage levels in the structure also increases, highlighting the significant impact of aftershocks on structural damage increments. Furthermore, we extend the proposed model to a bivariate seismic intensity measure and develop fragility surfaces. The proposed framework provides a novel approach and insights for tackling seismic fragility under multiple aftershocks.

Seismic Response of Vertical Hybrid Concrete/Steel Frames Considering Soil–Structure Interaction

The aim of this study is to investigate the seismic behavior of concrete/steel mixed structures. In engineering praxis, many buildings consist of two parts: one made of reinforced concrete and the other made of steel. There are several difficulties in the code-based seismic design of these structures due to the different dynamic responses of each discrete part. Seismic design codes, such as the IBC and Eurocode 8, do not provide instructions for structures consisting of two parts. In addition, they use a single-loading scenario, but there are many locations that are affected by more than one earthquake in a short period. Another drawback is that recent provisions do not consider soil–structure interaction effects. The specific issue addressed here is the seismic response of mixed structures, which is evaluated through inelastic time–history analysis. More specifically, the response indices involve height-wise distributions for peak interstory drift ratios, maximum floor horizontal displacements, maximum floor accelerations, and plastic hinge formations in the frame elements when they are subjected to seismic sequences of earthquakes, as well as in far fault ground motions for different soil types. The results reveal that sequential ground motions lead to increased displacement demands, and they affect the permanent displacements. This phenomenon appears in both cases of stiff and flexible soil, as well as for both regular and irregular frames. It is found that soil–structure interaction generally leads to lower values of IDR, and maximum horizontal displacement and acceleration in comparison with the case of rigid soil assumptions.

Seismic performance evaluation of reinforced concrete hilly buildings under sequence of earthquakes

SummaryThe present study investigates the effect of earthquake sequences on the response of reinforced concrete hilly buildings having typical configurations, that is, stepback and split‐foundation, with three different story ratios. A set of 30 as‐recorded mainshock‐aftershock sequence of earthquake ground motions is considered for this study. Mainshock acceleration time histories are scaled at two distinct intensity levels to obtain the mainshock‐damaged building. These mainshock‐damaged hilly buildings are then subjected to aftershocks. A comparative study is performed for various response quantities, such as peak interstory drift ratio, peak floor acceleration, and peak roof displacement of undamaged and mainshock‐damaged buildings under aftershocks. Further, fragility analysis is carried out to study the effect of aftershocks on undamaged and mainshock‐damaged hilly buildings. Subsequently, component‐wise seismic loss estimation due to damage in the non‐structural and structural components is performed. It is concluded from the study that the building components that contribute maximum to the expected repair cost ratio vary with respect to the intensity of the aftershocks. Also, the estimated seismic loss is higher in mainshock‐damaged split‐foundation buildings in comparison to stepback buildings.

Seismic collapse assessment of intermediate RC moment frames subjected to mainshock-aftershock sequences

This paper assessed the effects of mainshock-aftershock (MS-AS) sequences on the seismic vulnerability of intermediate reinforced concrete moment frames (IRCFs), considering various mainshock intensity scenarios. Moreover, given the limited library of real MS-AS ground motion sequences in the Iran plateau, a MS-AS sequence-selection method is proposed to generate artificial sequences that incorporate the magnitude, seismotectonic region, fault mechanism, and site condition of a target area. Three 4-, 7-, and 10-story IRCFs, modeled in OpenSees software, are subjected to real as well as artificial MS-AS sequences and collapse fragility curves are derived using the incremental dynamic analysis (IDA) results. The results show that the occurrence of mainshock can significantly increase the vulnerability of buildings during the aftershock. For various mainshock intensity scenarios, a general decrease in the collapse capacity is observed across all mainshock-damaged buildings, with a more significant decrease for taller buildings. Hence, it is essential to consider the effects of aftershock in the seismic design of IRCFs to mitigate the risk of collapse. Furthermore, by comparing the seismic responses caused by artificial sequences with those caused by real sequences, the validity of artificial sequence-selection method is confirmed; thus, artificial sequences can be used as a substitute for real sequences. Regression equations are also presented to predict the collapse capacity reduction of frames with different story numbers in presence of various mainshock intensity scenarios.

Response of embankment on liquefiable soil to sequential ground motions considering mitigation measures

The present study focused on gravel berm and stone column improvement techniques for mitigating the effects of liquefaction for an embankment resting on liquefiable ground. Three-dimensional finite element models are developed to evaluate the effectiveness of two countermeasure techniques for embankments resting on liquefiable soil. Three different embankment models are considered in this study as benchmark embankment model (BM), gravel berm embankment model (GBM), and stone column mitigation model (SCM). Foundation soil has been modeled using an elasto-plastic effective stress-based UBC3D-PLM model. Initially, the behaviour of three different embankment models (BM, GBM, and SCM) are evaluated under varying amplitude of cyclic input motion. Later, a seismic study is carried out considering 10 sequential earthquake motions to assess the effect of the aftershock following the main shock event. It can be seen that even small amplitude aftershocks can develop a high excess pore pressure ratio in the foundation soil, especially below the embankment toe. Moreover, a linear correlation has been observed between the input motion Arias intensity and the intensity at the embankment crest. In an overall observation, stone column mitigation was found to be a sound mitigation approach.

Effects of Aftershocks on the Seismic Performances of Reinforced Concrete Eccentric Frame Structures

Strong aftershocks have the potential to cause accumulated damage in structures, a feature which has been reported in post-earthquake reconnaissance studies, particularly for eccentric or irregular structures. This study aims to investigate the seismic behaviors of eccentric RC structural models under mainshock–aftershock (MSAS) sequences. In this study, three-dimensional structural models with eccentricities of 5%, 10%, 15%, 20%, 25%, and 30%, and an eccentricity of 0 (symmetric structural model) are developed by changing the positions of the centers of the structural mass. A static pushover analysis and a nonlinear time history analysis are conducted on the structural models with different eccentricities considering unidirectional and bidirectional earthquake loading (including mainshock ground motion and MSAS sequences). The amplitude of the aftershock ground motion is scaled according to the structural damage levels calibrated with the inter-story drift ratio (IDR). Furthermore, the differences in seismic responses between the unidirectional and bidirectional eccentric structures are discussed. The results show that the peak displacements of the unidirectional eccentric structures under MSAS sequences are nearly 1.4 times higher than those under mainshock ground motions. The structural seismic responses under unidirectional earthquake loading are more sensitive to the intensity of aftershock ground motions than those under bidirectional earthquake loading. Compared with unidirectional eccentric structures, bidirectional eccentric structures are more sensitive to the intensity of aftershock ground motions and have larger torsional angles and more complex displacement trends. The maximum displacement and the maximum IDR of bidirectional eccentric structures under MSAS sequences can reach 1.5 times and 1.4 times of those under mainshock ground motions, respectively.

Responses of a cantilever retaining wall under multiple earthquake sequences

The seismic behaviors of retaining walls were evaluated considering only a single earthquake. However, aftershock or foreshock ground motions can occasionally be as strong as mainshock ground motions and can aggravate damage to structures. This study investigated the seismic responses of a cantilever retaining wall under earthquake sequences by performing a series of finite-difference-based numerical simulations. We collected a suite of mainshock-aftershock and foreshock-mainshock sequential ground motions, which were used as the input ground motions in the numerical simulations. This study focused on revealing the differences in wall responses under a single earthquake and multiple earthquake sequences, considering a wide range of ground motion intensities. The relative wall movements caused by mainshock-aftershock and foreshock-mainshock sequential ground motions can be much larger than those caused by a single mainshock motion. In addition, the sequence itself had the effect of increasing the relative wall movements under strong sequential ground motions. However, the effects of the aftershock and foreshock motions on the wall responses were insignificant under weak sequential ground motions.

Seismic fragility assessment of aqueduct bent structures subjected to mainshock-aftershock sequences

This study aims to assess the seismic fragility of the aqueduct bent structure subjected to mainshock-aftershock (MS-AS) ground motion sequences by double incremental dynamic analysis method. The finite element model of the aqueduct bent structure is developed in OpenSees platform and numerical simulation technique is validated through physical test results. A series of nonlinear dynamic time-history analysis is carried out on the aqueduct bent structure to obtain its seismic response under different ground motion intensities. Furthermore, the incremental dynamic analysis (IDA) data cluster and median lines of the damage measure of the aqueduct bent structure are constructed. The advanced probabilistic seismic demand model (PSDM) of the aqueduct bent structure subjected to MS-AS ground motion sequences is proposed. Based on the proposed PSDM and the predefined damage states, the fragility surfaces of the aqueduct bent structure are also developed by adopting the vector-valued intensity measures considering the mainshock intensities (IMm) and aftershock intensities (IMa). Meanwhile, for the sake of comparation, the aftershock fragility curves of the aqueduct bent structure under different mainshock intensities are drawn, which emphasizes the importance of aftershocks in the seismic risk assessment. The analysis results indicate that the proposed piecewise probabilistic seismic demand model can better fit the logarithmic relationship between the aftershock intensities and the median of the reinforcement longitudinal tensile strain. The structural fragility probability increase with the increase of mainshock and aftershock intensities. It turns out that the mainshock intensities play a leading role in the damage probability of the aqueduct structure subjected to MS-AS ground motion sequences, the aftershock can cause additional damage to the aqueduct bent structure and reduce its seismic capacity. The meaningful results can provide supports for seismic optimization design, resilience improvement and post-earthquake repair-demand of the aqueduct structure in Southwest China.

Seismic Performance of Vertical Irregular Steel Frame Structure under Mainshock-Aftershock

Abstract: The majority of the structures are involved with architectural importance and it is highly impossible to achieve uniform structural properties in all directions. Hence earthquake resistant design codes considered it as irregular frames based on relative difference in the story properties. In many cases these irregularities are responsible for structural collapse of buildings under earthquake ground motions. The seismic response of buildings with irregular distribution of stiffness along the height may be differentfrom that ofregular building. Also, past earthquakes showed that structure may be subjected to sequence of ground motions but current codes do not have guidelines for such cases. It is considered that aftershock do not cause any more damage to damaged structure by the preceding mainshock ground motion. In this study steel moment resisting frame buildings are evaluated to understand the seismic response of vertical stiffness irregular frames subjected to mainshock-aftershock ground motions. The 9-story steel moment resisting frame building situated in Los Angeles is used in this study was originally developed as part of the SAC steel project. In this study soft and stiff storey case was considered at three different locations along the height, i.e., at bottom storey, mid-height storey and top storey. The single modification factor is used for irregularity. For comparison purpose dynamic properties of regular and irregular frames are kept same. Two sets of 6 records were selected representing a seismic hazard level of 2% and 10% probability of exceedance in 50 years respectively as mainshock. Two sets of 30 mainshockaftershock ground motion are considered for the study. These ground motions were developed using randomized approach. A non-linear time history analysis of regular and irregular building is carried out separately under mainshock and mainshockaftershock. The effect of building irregularity was studied for single storey modification at bottom storey, mid-storey and top storey with comparison to regular building. The comparison of a regular and irregular building is carried out in terms of maximum roof displacement and interstorey drift ratio. Also, comparison of frames under mainshock and mainshock-aftershock is done.

Study on seismic vulnerability analysis of the interaction system between saturated soft soil and subway station structures

The seismic vulnerability of interaction system of saturated soft soil and subway station structures was explored in this paper. The coupled nonlinear numerical models of interaction system were established using the u–p formulation of Biot′s theory to describe the saturated two-phase media. A refined finite element model of interaction system was developed to study its nonlinear seismic responses and seismic hazard mechanism. In this study, the multi-yield elastoplastic constitutive model was adopted for the soil, while a fiber section elastoplastic constitutive model was used for the structure. The seismic response of the structure was calculated by inputting the artificial seismic wave obtained from the power spectrum-triangular series method. The maximum inter-story drift angle was taken as a structural performance parameter for the subway station structure. The structural demand cloud was obtained under random ground motion sequences. Based on the probabilistic seismic demand model analysis method, the seismic vulnerability curve of the subway station structure was plotted, and the seismic vulnerability curve was analyzed as per the vulnerability of performance parameters. With the increase of soil strength, the vulnerability index of subway station structure under different peak acceleration ground motion decreased correspondingly. Based on the above vulnerability theory and analysis methods, it can be found from the above vulnerability theory and analysis methods that the subway station structure with established buried depth in saturated soft soil site exhibits a certain degree of safety and reliability, and can meet the seismic fortification goal of "no damage in small earthquakes, repairable in medium earthquakes and no collapse in large earthquakes". The results of vulnerability analysis are in line with the actual seismic survey, and the vulnerability analysis method proposed in this paper can be applied to the vulnerability analysis of underground structures on saturated soft soil foundation.

Energy-based procedures for seismic fragility analysis of mainshock-damaged buildings

In recent decades, significant research efforts have been devoted to developing fragility and vulnerability models for mainshock-damaged buildings, i.e., depending on the attained damage state after a mainshock ground motion (state-dependent fragility/vulnerability relationships). Displacement-based peak quantities, such as the maximum interstory drift ratio, are widely adopted in fragility analysis to define both engineering demands and structural capacities at the global and/or local levels. However, when considering ground-motion sequences, the use of peak quantities may lead to statistical inconsistencies (e.g., fragility curves’ crossings) due to inadequate consideration of damage accumulation. In this context, energy-based engineering demand parameters (EDPs), explicitly accounting for cumulative damage, can help address this issue. This paper provides an overview of recent findings on the development of aftershock-fragility models of mainshock-damaged buildings. Particular focus is given to state-of-the-art frameworks for fragility analyses based on cumulative damage parameters. Moreover, a literature review on damage indices and energy-based concepts and approaches in earthquake engineering is reported to better understand the main advantages of the mostly adopted energy-based parameters, as well as their limitations. Different refinement levels of seismic response analyses to derive fragility relationships of mainshock-damaged buildings are also discussed. Finally, the benefits of adopting energy-based EDPs rather than, or in addition to, peak quantities in state-dependent fragility analyses are demonstrated on a reinforced concrete frame building. Specifically, a refined lumped plasticity modeling approach is adopted, and sequential cloud-based time-history analyses of a Multi-Degree-of-Freedom (MDoF) model are carried out. The results highlight that energy-based approaches for fragility analysis effectively capture damage accumulation during earthquake sequences without inconsistencies in the obtained statistical models. On the other hand, estimating global or local structural capacity in terms of cumulative EDPs is still challenging. Further experimental data are needed to better calibrate the quantification of energy-based damaged states.

A performance‐based seismic loading protocol: The generated sequential ground motion

AbstractA realistic performance‐based seismic loading protocol called generated sequential ground motion (GSGM) has been developed in this paper. GSGM is a ground motion fabricated from segments of real recorded ground motions that could enable the introduction of performance‐based seismic assessment and design to experimental testing in setups such as shaking table testing. It can also significantly reduce the number of nonlinear time history analyses required in performance‐based seismic design. The protocol optimizes the behavioral information output of an experimental test or numerical analysis by incorporating dynamic demands corresponding to design limit states with different probabilities of exceedance (i.e. 10%, 5%, and 2% in 50 years) in a single record. In addition, since the segments are matched to relevant target spectra, the number of ground motions required to estimate the mean response is reduced. This paper presents the algorithm developed to produce the GSGM. The capability of the GSGM to replicate the structural responses produced by code‐compliant suites, and a suite of 100 ground motions as a more robust estimation of the actual response is investigated. The results of the case study bridge pier show that the drift variation of the GSGMs compared to code‐compliant suites is within 10%. Compared to the estimate of the actual response, the drift variation of GSGMs and the code‐compliant suites is 20% and 15%, respectively, and the damage variation is 30% and 15%, respectively. Furthermore, considering other relevant intensity measures when producing GSGMs can reduce these variations. This study suggests that the GSGM can replicate structural responses of the current code procedures.

Seismic evaluation of vertically irregular RC frames subjected to mainshock-aftershock sequences of near-fault and far-fault ground motions

Sequential ground motions can significantly affect the seismic demands of reinforced concrete (RC) frames with irregularity along the height. In this paper, the effect of sequential near-fault and far-fault ground motion records with strong and weak aftershocks on the seismic demands of vertically irregular 10-story RC building frames is investigated. The vertically irregular frames with stiffness and mass irregularities were created by applying a modification factor of MF = 2 in three different locations along the height of the reference regular frame. Eigenvalue analysis was performed to evaluate the dynamic characteristics of the structures (i.e. the period and the effective modal participating mass ratio for different modes of vibration). In order to study the seismic demands of these structures, the non-linear response history analysis (NL-RHA), which is the most accurate method of seismic evaluation of structures, was carried out. The floor displacements, story drifts, plastic hinge rotations and residual story drifts obtained from the NL-RHAs were exhaustively studied, and the effect of seismic sequences on the seismic responses was investigated. Furthermore, the enhanced pushover analyses, including the modal pushover analysis (MPA), consecutive modal pushover (CMP) procedure, extended N2 (EN2) method, single-run multi-mode pushover (SMP) method and non-adaptive displacement-based pushover (NADP) procedure were implemented to consider the effect of higher modes in computing the seismic demands of the structures subjected to sequential near-fault and far-fault records as well as mainshocks. The results accentuate that the maximum seismic demands and damage of the structure for sequential near-fault and far-fault records with strong aftershocks are much greater than those for the corresponding mainshocks. Moreover, the enhanced pushover analyses in the case of sequential records have usually less accuracy in estimating the seismic demands in comparison with mainshocks only.

3D numerical characterization of a dissipative connection system for retrofit of prefabricated existing RC sheds

Prefabricated industrial sheds featured a high seismic vulnerability during the 2012 Emilia earthquake (Italy). The buildings typically exhibited a rigid collapse mechanism that was a consequence of the loss of support between columns, beams and roof elements.The study presents a numerical characterization of a novel dissipative connection system (DCS) designed to improve the seismic performance of industrial sheds. The device, which is placed on the top of the columns, exploits the movement of a rigid body on a sloped surface to provide horizontal stiffness and control the lateral displacement of the beam.A 3D finite element model of the prototype is formulated in Abaqus and used to switch the backbone curve from the scaled model to the full-scale device. A parametric study is conducted to evaluate the influence of the slope of the contact surface and the coefficient of friction on the output force of the system.In the second part of the study, non-linear dynamic analyses are performed on a finite element model of a portal frame implementing, at beam-column joints, either the DCS or a pure friction connection. The results highlight the effectiveness of the DCS in controlling beam-to-column displacements, reducing shear forces on the top of columns, and limiting residual displacements that can accrue during ground motion sequences.

Performance of RC Frames with Mass and Stiffness Irregularity under Sequential Ground Motion

The seismic design codes consider only single earthquake events, ignoring the possibility of multiple earthquakes. In case a structure suffers damage during first event and prior to its repair more aftershocks occur the structure may collapse. An elaborative directive study on the inelastic seismic response of 11storey RC (2-D) moment-resisting frame having mass irregularity at sixth floor, and stiffness irregularity in sixth story has been done has been done. The frame has been examined under a sequence of a main shock followed by six aftershocks. Time history synthetically generated from the design response spectrum explained in IS-1893 Pt II: 2016 has been taken for non linear dynamic analysis with the help of SAP 2000 software. A comparison is done to understand the inelastic response of the frame taking into account the location of plastic hinges, maximum top displacement, residual top displacement, and storey drift. It is seen that the structure with the stiffness irregularity is the most vulnerable amongst all.

Exploring the performance of experimentally benchmarked RC bridge pier models when subjected to sequential seismic shocks

In this paper we explore the performance of RC bridge piers to seismic ground motion sequences, using both experimental and numerical models. Four RC columns were tested on the University of Bristol’s shake table. These columns contained both well-confined and poor-confined cases. Spectrally matched by near-field without pulse (NFWP), near-field pulse-like (NFPL) and far-field (FF) ground motion records where employed in a sequential/progressive fashion ranging from (I) slight damage (II) extensive damage (III) complete damage and (IV) aftershock cases. These experimental test results are then used to develop a benchmarked OpenSees model of this bridge pier. The importance of the concrete tension constitutive model is highlighted. The differences between sequential (progressive damage) and neglecting sequential seismic events are discussed. The benchmarked model is then used for a heuristic case using incremental dynamic analyses. A comparison is made between drift and energy dissipation performance measures, that suggests drift cannot identify the increased system damage induced by sequential events.

Vulnerability assessment and collapse simulation of unreinforced masonry structures subjected to sequential ground motions

Vulnerability assessment and collapse simulation of unreinforced masonry structures subjected to sequential ground motions

A generalized ground-motion model for consistent mainshock–aftershock intensity measures using successive recurrent neural networks

Several recent studies have investigated the risk posed to structures by earthquake sequences, proposing state-dependent fragility/vulnerability models for assets in damaged conditions. However, a critical component for such efforts, i.e., ground-motion record selection, has received relatively minor consideration. Specifically, utilization of “consistent” mainshock (MS)–aftershock (AS) ground motions is desirable in practical applications. Such consistency in selecting MS–AS sequences requires proper consideration of the correlations between and within the intensity measures of MS and AS ground motions. Most of the studies in this domain utilize spectral accelerations as the considered ground-motion intensity measures and rely on empirical linear correlation models between the intensity measures of MS and AS ground motions for developing, for instance, record selection approaches. This study proposes a generalized ground-motion model (GGMM) to estimate consistent 30 × 1 vectors of intensity measures for mainshocks (denoted as IMMS) and aftershocks (denoted as IMAS) using a framework of successive long-short-term-memory (LSTM) recurrent neural network (RNN). The vectors of IMMS and IMAS consist of geometric means of significant duration (D_{5 - 95,geom}), Arias intensity (I_{a,geom}), cumulative absolute velocity (CAV_{geom}), peak ground velocity (PGV_{geom}), peak ground acceleration (PGA_{geom}) and RotD50 spectral acceleration (S_{a} left( T right)) at 25 periods for both MS and AS ground motions. The proposed RNN-based GGMM is trained on a carefully selected set of ~ 700 crustal and subduction recorded MS–AS sequences. The inputs to the framework include a 5 × 1 vector of source and site parameters for MS and AS recordings. The residuals of the trained LSTM-based RNN are further used to develop empirical covariance structures for IMMS and IMAS. The proposed framework is finally illustrated to select MS–AS ground motions based on IMMS and IMAS using a multi-criteria objective function. The selected MS–AS ground motion sequences are then used to perform non-linear time-history analyses of a case-study two-spanned symmetric bridge structure. The obtained engineering demand parameters are evaluated and critically discussed.