Pub Date : 2026-02-12DOI: 10.1016/j.engstruct.2026.122312
Hermawan Sutejo , Yu-Chen Ou
Buckling of longitudinal reinforcement in compression, often followed by low-cycle fatigue fracture in tension, is a primary mechanism driving strength degradation in reinforced concrete flexural members subjected to large-displacement reversals. This study proposes a mechanics-based analytical model to predict the buckling length of longitudinal reinforcement restrained by rectilinear transverse reinforcement. The model captures buckling lengths over a non-integer interval of tie spacings by incorporating end transition regions beyond the outermost hoops bounding the buckling region. The buckling-restraint stiffness is formulated by combining axial and bending components. The axial component is adjusted to reflect hoop type and hook-bent angle through a geometric effectiveness factor, while the bending component is evaluated from the flexural response of transverse ties. The model is validated using 38 beam and 32 column specimens collected from the literature. The proposed model achieves improved accuracy relative to the models by Su et al. and Dhakal & Maekawa, with average prediction errors of 6.5 % for beams and 10.1 % for columns, compared to 9.8 % and 12.3 % for Su et al. and 26.0 % and 22.3 % for Dhakal & Maekawa, respectively. Parametric reanalysis shows that excluding either the axial reduction factor or the bending component increases the error by about 20 %, and neglecting both increases the error by up to 55 %, demonstrating that both mechanisms are essential for reliable buckling-length prediction.
{"title":"Prediction model for longitudinal reinforcement buckling in reinforced concrete beams and columns with rectilinear hoops","authors":"Hermawan Sutejo , Yu-Chen Ou","doi":"10.1016/j.engstruct.2026.122312","DOIUrl":"10.1016/j.engstruct.2026.122312","url":null,"abstract":"<div><div>Buckling of longitudinal reinforcement in compression, often followed by low-cycle fatigue fracture in tension, is a primary mechanism driving strength degradation in reinforced concrete flexural members subjected to large-displacement reversals. This study proposes a mechanics-based analytical model to predict the buckling length of longitudinal reinforcement restrained by rectilinear transverse reinforcement. The model captures buckling lengths over a non-integer interval of tie spacings by incorporating end transition regions beyond the outermost hoops bounding the buckling region. The buckling-restraint stiffness is formulated by combining axial and bending components. The axial component is adjusted to reflect hoop type and hook-bent angle through a geometric effectiveness factor, while the bending component is evaluated from the flexural response of transverse ties. The model is validated using 38 beam and 32 column specimens collected from the literature. The proposed model achieves improved accuracy relative to the models by Su et al. and Dhakal & Maekawa, with average prediction errors of 6.5 % for beams and 10.1 % for columns, compared to 9.8 % and 12.3 % for Su et al. and 26.0 % and 22.3 % for Dhakal & Maekawa, respectively. Parametric reanalysis shows that excluding either the axial reduction factor or the bending component increases the error by about 20 %, and neglecting both increases the error by up to 55 %, demonstrating that both mechanisms are essential for reliable buckling-length prediction.</div></div>","PeriodicalId":11763,"journal":{"name":"Engineering Structures","volume":"353 ","pages":"Article 122312"},"PeriodicalIF":6.4,"publicationDate":"2026-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146154156","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-12DOI: 10.1016/j.engstruct.2026.122325
Trevor Zhiqing Yeow , Koichi Kusunoki
In structural health monitoring, correct classification of a building’s inelastic deformation mode (i.e., total-yield or soft-story) is needed for accurate safety evaluations. However, sensors are usually not placed on all floors in most applications, making inelastic deformation mode classification difficult. In this study, features based on plastic displacements and sensor location are proposed for training and evaluating inelastic deformation mode classification models. The importance of the newly proposed features was compared against other features proposed in literature based on peak floor acceleration and velocity response, cumulative absolute velocity and jerk. A large building response database was created from numerical simulations of a wide range of reinforced concrete frame structures exhibiting different inelastic deformation modes for evaluating feature importance. It was found that the newly proposed features ranked highly when applying the Minimum Redundancy Maximum Relevancy algorithm to the response database compared to past features. Furthermore, a k-Nearest Neighbor classification model trained using a feature set containing the proposed features and building-level ductility response resulted in a more accurate model compared to only using existing features (misclassification rate of 10 % versus 29 %). These results demonstrate the suitability of the proposed features for training and evaluating building inelastic deformation mode classification models.
{"title":"Plastic displacement features for classifying the inelastic deformation mode of instrumented buildings with few sensors considering sensor locations","authors":"Trevor Zhiqing Yeow , Koichi Kusunoki","doi":"10.1016/j.engstruct.2026.122325","DOIUrl":"10.1016/j.engstruct.2026.122325","url":null,"abstract":"<div><div>In structural health monitoring, correct classification of a building’s inelastic deformation mode (i.e., total-yield or soft-story) is needed for accurate safety evaluations. However, sensors are usually not placed on all floors in most applications, making inelastic deformation mode classification difficult. In this study, features based on plastic displacements and sensor location are proposed for training and evaluating inelastic deformation mode classification models. The importance of the newly proposed features was compared against other features proposed in literature based on peak floor acceleration and velocity response, cumulative absolute velocity and jerk. A large building response database was created from numerical simulations of a wide range of reinforced concrete frame structures exhibiting different inelastic deformation modes for evaluating feature importance. It was found that the newly proposed features ranked highly when applying the Minimum Redundancy Maximum Relevancy algorithm to the response database compared to past features. Furthermore, a <em>k</em>-Nearest Neighbor classification model trained using a feature set containing the proposed features and building-level ductility response resulted in a more accurate model compared to only using existing features (misclassification rate of 10 % versus 29 %). These results demonstrate the suitability of the proposed features for training and evaluating building inelastic deformation mode classification models.</div></div>","PeriodicalId":11763,"journal":{"name":"Engineering Structures","volume":"353 ","pages":"Article 122325"},"PeriodicalIF":6.4,"publicationDate":"2026-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146154153","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-12DOI: 10.1016/j.engstruct.2026.122330
Zhu-Yu Sun , Yu-Tao Guo , Kang Ge , Chao Hou , Zhen-Zhong Hu
Offshore bridges operate in complex ocean environments, making structural analysis, design, and monitoring more challenging. Existing typical time history analysis and nonlinear model updating based on finite-element methods are computationally intensive and time consuming, limiting the usage in many scenarios. To create a more efficient analytical tool, a Deep Learning-based Offshore Bridge Predictor (DeepOBP) is proposed. The model integrates structural characteristics and coupled dynamic loads in the ocean environments, enabling millisecond level and high precision nonlinear dynamic offshore bridge response predictions. A differentiable structural inverse framework (Inverse DeepOBP) couples the surrogate model with gradient-based optimization is further developed to enable rapid damage identification and model calibration for structural health monitoring. The experimental results show that DeepOBP demonstrates high accuracy under both normal operating conditions and multihazard coupled conditions, with R² = 0.93 and 0.92. Inverse DeepOBP delivers more than a 10-fold and more than a 104-fold speed-up over surrogate-based model updating with the heuristic algorithm and nonlinear finite element model updating, respectively, while maintaining relative errors below 7 % for each identified parameter, enabling efficient structural analyses and real-time monitoring.
{"title":"Deep learning-based realtime multiload response prediction and inverse analysis of offshore bridges","authors":"Zhu-Yu Sun , Yu-Tao Guo , Kang Ge , Chao Hou , Zhen-Zhong Hu","doi":"10.1016/j.engstruct.2026.122330","DOIUrl":"10.1016/j.engstruct.2026.122330","url":null,"abstract":"<div><div>Offshore bridges operate in complex ocean environments, making structural analysis, design, and monitoring more challenging. Existing typical time history analysis and nonlinear model updating based on finite-element methods are computationally intensive and time consuming, limiting the usage in many scenarios. To create a more efficient analytical tool, a Deep Learning-based Offshore Bridge Predictor (DeepOBP) is proposed. The model integrates structural characteristics and coupled dynamic loads in the ocean environments, enabling millisecond level and high precision nonlinear dynamic offshore bridge response predictions. A differentiable structural inverse framework (Inverse DeepOBP) couples the surrogate model with gradient-based optimization is further developed to enable rapid damage identification and model calibration for structural health monitoring. The experimental results show that DeepOBP demonstrates high accuracy under both normal operating conditions and multihazard coupled conditions, with R² = 0.93 and 0.92. Inverse DeepOBP delivers more than a 10-fold and more than a 10<sup>4</sup>-fold speed-up over surrogate-based model updating with the heuristic algorithm and nonlinear finite element model updating, respectively, while maintaining relative errors below 7 % for each identified parameter, enabling efficient structural analyses and real-time monitoring.</div></div>","PeriodicalId":11763,"journal":{"name":"Engineering Structures","volume":"353 ","pages":"Article 122330"},"PeriodicalIF":6.4,"publicationDate":"2026-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146154154","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-12DOI: 10.1016/j.engstruct.2026.122339
Yirui Sun, Yujie Chen, Zonghan Xie
Advances in modelling and simulation are driving innovation in mechanical joint design. However, the lack of standardized evaluation criteria hinders meaningful comparison across geometries, rendering the rational design and improvement difficult. To address this, we studied three representative joint shapes—trapezoid, circle, and ellipse. Finite element analysis (FEA) was employed to evaluate their tensile performance within the elastic regime. The elliptical joint showed the highest stiffness, while the circular joint exhibited the greatest load capability and resilience. Joint performance was also influenced by friction coefficient, yield strength, and blade number. Applying edge constraints notably enhanced performance, especially for single-blade joints, with up to 7.6 × increase in load capability and 5.4 × in resilience for circular joints, and 11.2 × in stiffness for trapezoidal joints. An Ashby-type plot was developed to support the comparative selection of joint designs. These results provide a foundation for establishing standardized evaluation criteria for tensile joint performance.
{"title":"Computational analysis of interlocking joints with different geometries under tensile loads","authors":"Yirui Sun, Yujie Chen, Zonghan Xie","doi":"10.1016/j.engstruct.2026.122339","DOIUrl":"10.1016/j.engstruct.2026.122339","url":null,"abstract":"<div><div>Advances in modelling and simulation are driving innovation in mechanical joint design. However, the lack of standardized evaluation criteria hinders meaningful comparison across geometries, rendering the rational design and improvement difficult. To address this, we studied three representative joint shapes—trapezoid, circle, and ellipse. Finite element analysis (FEA) was employed to evaluate their tensile performance within the elastic regime. The elliptical joint showed the highest stiffness, while the circular joint exhibited the greatest load capability and resilience. Joint performance was also influenced by friction coefficient, yield strength, and blade number. Applying edge constraints notably enhanced performance, especially for single-blade joints, with up to 7.6 × increase in load capability and 5.4 × in resilience for circular joints, and 11.2 × in stiffness for trapezoidal joints. An Ashby-type plot was developed to support the comparative selection of joint designs. These results provide a foundation for establishing standardized evaluation criteria for tensile joint performance.</div></div>","PeriodicalId":11763,"journal":{"name":"Engineering Structures","volume":"353 ","pages":"Article 122339"},"PeriodicalIF":6.4,"publicationDate":"2026-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146154152","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-12DOI: 10.1016/j.engstruct.2026.122289
Rafał Walczak , Wit Derkowski
This study examines the structural behaviour of aged prefabricated post-tensioned concrete beams subjected to simulated anchorage failure - an essential aspect in assessing mid-20th-century industrial structures. Full-scale crane girders, in service for over 50 years, were tested to replicate emergency scenarios involving partial loss of tendon anchorages. The investigation addressed various anchorage failure configurations, grout quality levels, shear span-to-depth ratios (a/d), and the impact of low transverse reinforcement. Results showed that even in severe cases - such as loss of both bottom anchorages and insufficient tendon grouting - the beams did not exhibit brittle behaviour. Clear warning symptoms like large deflections and visible cracking preceded failure, although the capacity dropped by up to 50 %. In contrast, the failure of top tendon anchorage had a negligible impact on load-bearing capacity. Beams with low shear slenderness demonstrated higher ultimate strength, typically failing through concrete crushing, while more slender beams followed beam-type failure modes. However, anchorage failures may result in a distinct failure mode of the member. Numerical simulations using DIANA FEA, validated against the test results, extended the analysis to additional damage scenarios. Notably, simulation of the failure of all four bottom anchorages, out of five tendons in the beam, indicated that the beam could sustain load only until initial cracking, after which brittle failure occurred - identifying a critical threshold for safety evaluations. Despite limited stirrup reinforcement, all beams demonstrated sufficient shear performance. These findings contribute valuable insight into the structural assessment and sustainable long-term use or reuse of ageing post-tensioned elements, supporting more informed and sustainable infrastructure decisions.
{"title":"When anchorage fails: Assessing old post-tensioned precast beams in service","authors":"Rafał Walczak , Wit Derkowski","doi":"10.1016/j.engstruct.2026.122289","DOIUrl":"10.1016/j.engstruct.2026.122289","url":null,"abstract":"<div><div>This study examines the structural behaviour of aged prefabricated post-tensioned concrete beams subjected to simulated anchorage failure - an essential aspect in assessing mid-20th-century industrial structures. Full-scale crane girders, in service for over 50 years, were tested to replicate emergency scenarios involving partial loss of tendon anchorages. The investigation addressed various anchorage failure configurations, grout quality levels, shear span-to-depth ratios (a/d), and the impact of low transverse reinforcement. Results showed that even in severe cases - such as loss of both bottom anchorages and insufficient tendon grouting - the beams did not exhibit brittle behaviour. Clear warning symptoms like large deflections and visible cracking preceded failure, although the capacity dropped by up to 50 %. In contrast, the failure of top tendon anchorage had a negligible impact on load-bearing capacity. Beams with low shear slenderness demonstrated higher ultimate strength, typically failing through concrete crushing, while more slender beams followed beam-type failure modes. However, anchorage failures may result in a distinct failure mode of the member. Numerical simulations using DIANA FEA, validated against the test results, extended the analysis to additional damage scenarios. Notably, simulation of the failure of all four bottom anchorages, out of five tendons in the beam, indicated that the beam could sustain load only until initial cracking, after which brittle failure occurred - identifying a critical threshold for safety evaluations. Despite limited stirrup reinforcement, all beams demonstrated sufficient shear performance. These findings contribute valuable insight into the structural assessment and sustainable long-term use or reuse of ageing post-tensioned elements, supporting more informed and sustainable infrastructure decisions.</div></div>","PeriodicalId":11763,"journal":{"name":"Engineering Structures","volume":"353 ","pages":"Article 122289"},"PeriodicalIF":6.4,"publicationDate":"2026-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146154155","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-11DOI: 10.1016/j.engstruct.2026.122345
Delbaz Samadian, Annalisa Occhipinti, Imrose B. Muhit, Nashwan Dawood
Accurately assessing the vulnerability of critical building portfolios is fundamental for regional risk management and decision support, especially in regions facing sequential earthquake–flood events exacerbated by climate change. Such compound disasters pose severe environmental challenges, yet current practice lacks reliable surrogate models to rapidly predict structural response and damage-relevant demand parameters under combined seismic and flood loading. This study addresses that gap by introducing a soft computing approach, the Stacked Attention-based Long Short-Term Memory network (Stack-AttenLSTM), to efficiently predict key structural response quantities under sequential earthquake–flood hazards. In this framework, structural vulnerability is interpreted in a performance-based sense, whereby hazard-induced response metrics serve as proxies for damage susceptibility rather than direct loss estimation. The surrogate model predicts key engineering demand parameters (EDPs), including the maximum inter-storey drift ratio (MIDR), maximum floor acceleration (MFA), and maximum base shear (MBS), which are widely used indicators of structural damage and vulnerability. To develop this model, a large-scale meta-database comprising 30,000 steel special moment-resisting frame (SMRF) buildings is first generated to capture structural variability across low-, mid-, and high-rise typologies, from which a representative subset is selected for detailed high-fidelity three-dimensional (3D) nonlinear time-history analyses (NLTHA) under sequential earthquake–flood loading and surrogate model training. Flood loading is represented using computational fluid dynamics (CFD) simulations with a dam-break–type inflow condition, employed as a conservative hydrodynamic proxy to study flow-induced forces under extreme inundation scenarios. Multiple Stack-AttenLSTM architectures are trained and evaluated, and the final model is selected for its optimal balance of predictive accuracy and computational efficiency, enabling rapid yet reliable response prediction. The proposed model achieves high predictive accuracy, with coefficients of determination (R²) approaching 0.88 and low error metrics across all hazard scenarios, demonstrating its effectiveness for rapid multi-hazard vulnerability assessment. Although explicit fragility or loss models are not derived, the Stack-AttenLSTM framework is suitable for integration with early warning systems and digital twin platforms, enabling real-time monitoring, improved uncertainty management, and proactive disaster response.
{"title":"Stack-AttenLSTM: A surrogate deep learning model for sequential earthquake-flood structural response assessment of steel buildings","authors":"Delbaz Samadian, Annalisa Occhipinti, Imrose B. Muhit, Nashwan Dawood","doi":"10.1016/j.engstruct.2026.122345","DOIUrl":"10.1016/j.engstruct.2026.122345","url":null,"abstract":"<div><div>Accurately assessing the vulnerability of critical building portfolios is fundamental for regional risk management and decision support, especially in regions facing sequential earthquake–flood events exacerbated by climate change. Such compound disasters pose severe environmental challenges, yet current practice lacks reliable surrogate models to rapidly predict structural response and damage-relevant demand parameters under combined seismic and flood loading. This study addresses that gap by introducing a soft computing approach, the Stacked Attention-based Long Short-Term Memory network (Stack-AttenLSTM), to efficiently predict key structural response quantities under sequential earthquake–flood hazards. In this framework, structural vulnerability is interpreted in a performance-based sense, whereby hazard-induced response metrics serve as proxies for damage susceptibility rather than direct loss estimation. The surrogate model predicts key engineering demand parameters (EDPs), including the maximum inter-storey drift ratio (MIDR), maximum floor acceleration (MFA), and maximum base shear (MBS), which are widely used indicators of structural damage and vulnerability. To develop this model, a large-scale meta-database comprising 30,000 steel special moment-resisting frame (SMRF) buildings is first generated to capture structural variability across low-, mid-, and high-rise typologies, from which a representative subset is selected for detailed high-fidelity three-dimensional (3D) nonlinear time-history analyses (NLTHA) under sequential earthquake–flood loading and surrogate model training. Flood loading is represented using computational fluid dynamics (CFD) simulations with a dam-break–type inflow condition, employed as a conservative hydrodynamic proxy to study flow-induced forces under extreme inundation scenarios. Multiple Stack-AttenLSTM architectures are trained and evaluated, and the final model is selected for its optimal balance of predictive accuracy and computational efficiency, enabling rapid yet reliable response prediction. The proposed model achieves high predictive accuracy, with coefficients of determination (R²) approaching 0.88 and low error metrics across all hazard scenarios, demonstrating its effectiveness for rapid multi-hazard vulnerability assessment. Although explicit fragility or loss models are not derived, the Stack-AttenLSTM framework is suitable for integration with early warning systems and digital twin platforms, enabling real-time monitoring, improved uncertainty management, and proactive disaster response.</div></div>","PeriodicalId":11763,"journal":{"name":"Engineering Structures","volume":"353 ","pages":"Article 122345"},"PeriodicalIF":6.4,"publicationDate":"2026-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146154150","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-11DOI: 10.1016/j.engstruct.2026.122338
Yansong Liu , Meng Zou , Yingchun Qi , Ruizhe Wu , Jiangquan Li , Jiafeng Song , Shucai Xu , Weiguang Fan , Qingyu Yu
The prediction of deformation patterns and full-field stress responses in thin-walled tubes driven by unit cell images remains largely unexplored. Two major challenges exist: how to directly construct finite element models of thin-walled tubes from 2D unit cell images to enable structural response simulation, and how to achieve multimodal, temporal response prediction based on a single image. To address these issues, this study proposes an integrated image-driven prediction framework that fuses a generative adversarial network with a temporal modeling network, enabling direct generation of 10-frame stress evolution sequences under axial compression from a static unit cell image. To support data-driven modeling, we developed a highly automated simulation platform, which streamlines the entire pipeline from image-based structure generation to automated modeling and finite element simulation, allowing for the construction of a large-scale image-to-stress dataset. Experimental results demonstrate that the proposed fusion model improves the average peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) on the test set by 14.70 % and 5.68 %, respectively, compared to the original Pix2Pix model, while maintaining an average inference time of only 0.0288 s per image, highlighting both accuracy and efficiency. Moreover, the model exhibits strong robustness across key metrics such as stress area ratio, Hausdorff boundary distance, and high-stress region error. On previously unseen test configurations, the average relative error of the predicted mean stress is approximately 4.65 %. This study presents a highly efficient and scalable paradigm for full-field response modeling and rapid performance prediction of complex structures in an image-driven manner.
{"title":"Image-driven multimodal prediction of deformation and stress evolution in thin-walled structures","authors":"Yansong Liu , Meng Zou , Yingchun Qi , Ruizhe Wu , Jiangquan Li , Jiafeng Song , Shucai Xu , Weiguang Fan , Qingyu Yu","doi":"10.1016/j.engstruct.2026.122338","DOIUrl":"10.1016/j.engstruct.2026.122338","url":null,"abstract":"<div><div>The prediction of deformation patterns and full-field stress responses in thin-walled tubes driven by unit cell images remains largely unexplored. Two major challenges exist: how to directly construct finite element models of thin-walled tubes from 2D unit cell images to enable structural response simulation, and how to achieve multimodal, temporal response prediction based on a single image. To address these issues, this study proposes an integrated image-driven prediction framework that fuses a generative adversarial network with a temporal modeling network, enabling direct generation of 10-frame stress evolution sequences under axial compression from a static unit cell image. To support data-driven modeling, we developed a highly automated simulation platform, which streamlines the entire pipeline from image-based structure generation to automated modeling and finite element simulation, allowing for the construction of a large-scale image-to-stress dataset. Experimental results demonstrate that the proposed fusion model improves the average peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) on the test set by 14.70 % and 5.68 %, respectively, compared to the original Pix2Pix model, while maintaining an average inference time of only 0.0288 s per image, highlighting both accuracy and efficiency. Moreover, the model exhibits strong robustness across key metrics such as stress area ratio, Hausdorff boundary distance, and high-stress region error. On previously unseen test configurations, the average relative error of the predicted mean stress is approximately 4.65 %. This study presents a highly efficient and scalable paradigm for full-field response modeling and rapid performance prediction of complex structures in an image-driven manner.</div></div>","PeriodicalId":11763,"journal":{"name":"Engineering Structures","volume":"353 ","pages":"Article 122338"},"PeriodicalIF":6.4,"publicationDate":"2026-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146153740","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-11DOI: 10.1016/j.engstruct.2026.122326
Jinchao Gu , Xiongyan Li , Wei Wang , Wenfeng Du , Zhuang Xia , Suduo Xue
Bolted spherical joints (BSJs) are critical components in spatial grid structures. Traditionally, they consist of solid steel spheres and bolt holes, offering good manufacturability and versatility but limited potential for weight reduction and performance enhancement. This study proposes an intelligent lightweight design method that integrates topology optimization (TO) and deep learning. A design database was first established using the SIMP method. A three-dimensional Least Squares Generative Adversarial Network (3D-LSGAN) was then trained to generate innovative structural designs. After post-processing, the performance of these designs was evaluated and optimized using a TOPSIS-based multi-attribute decision-making approach. Validation on a double-layer grid structure shows that the optimal intelligent design maintains acceptable maximum von Mises stress levels across multiple loading conditions, significantly improves stress distribution uniformity, and achieves mass reductions of 83.86 % and 5.46 % compared with the initial and single-TO designs, respectively. These results demonstrate that the proposed framework provides an effective approach for the intelligent and lightweight design of BSJs.
{"title":"Intelligent lightweight design of bolted spherical joints in spatial grid structures based on topology optimization with 3D-LSGAN predictor","authors":"Jinchao Gu , Xiongyan Li , Wei Wang , Wenfeng Du , Zhuang Xia , Suduo Xue","doi":"10.1016/j.engstruct.2026.122326","DOIUrl":"10.1016/j.engstruct.2026.122326","url":null,"abstract":"<div><div>Bolted spherical joints (BSJs) are critical components in spatial grid structures. Traditionally, they consist of solid steel spheres and bolt holes, offering good manufacturability and versatility but limited potential for weight reduction and performance enhancement. This study proposes an intelligent lightweight design method that integrates topology optimization (TO) and deep learning. A design database was first established using the SIMP method. A three-dimensional Least Squares Generative Adversarial Network (3D-LSGAN) was then trained to generate innovative structural designs. After post-processing, the performance of these designs was evaluated and optimized using a TOPSIS-based multi-attribute decision-making approach. Validation on a double-layer grid structure shows that the optimal intelligent design maintains acceptable maximum von Mises stress levels across multiple loading conditions, significantly improves stress distribution uniformity, and achieves mass reductions of 83.86 % and 5.46 % compared with the initial and single-TO designs, respectively. These results demonstrate that the proposed framework provides an effective approach for the intelligent and lightweight design of BSJs.</div></div>","PeriodicalId":11763,"journal":{"name":"Engineering Structures","volume":"353 ","pages":"Article 122326"},"PeriodicalIF":6.4,"publicationDate":"2026-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146154151","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1016/j.engstruct.2026.122270
Amir Hossein Asjodi
This paper employs supervised and unsupervised learning methods to present hazard-based seismic fragility functions for Steel Moment-Resisting Frame (SMRF) buildings. The database supporting this research comprises structural responses of over 12,000 time history analyses for 100 SMRF buildings lumped into three categories: low-, mid-, and high-rise. The ground motions have been selected to represent three hazard levels, resulting in Service Level Earthquake (SLE), Design Basis Earthquake (DBE), and Maximum Considered Earthquake (MCE). Considering the primary period of each building and the target response spectra, a set of ground motions is selected, and the peak story drift ratios are extracted. Subsequently, unsupervised clustering techniques are employed to identify drift thresholds that distinguish between different damage states across various hazard levels, thereby refining the fixed boundaries recommended in existing codes and guidelines. Supervised learning techniques, on the other hand, are employed to predict the maximum drift ratio using features from ground motions and structural periods. The resulting drift ratio serves as an Engineering Demand Parameter (EDP), which, along with the hazard-informed drift threshold, is used to generate a machine learning-based fragility function. The proposed approach enables damage state identification of SMRF buildings under a specific ground motion, using only structural periods and signal features, without requiring detailed structural response data. The results of this study provide a set of site-specific hazard-based fragility curves, supporting seismic risk and loss assessment across different earthquake intensities.
{"title":"Hazard-based seismic fragility functions for steel moment-resisting frame buildings through data-driven damage state identification","authors":"Amir Hossein Asjodi","doi":"10.1016/j.engstruct.2026.122270","DOIUrl":"10.1016/j.engstruct.2026.122270","url":null,"abstract":"<div><div>This paper employs supervised and unsupervised learning methods to present hazard-based seismic fragility functions for Steel Moment-Resisting Frame (SMRF) buildings. The database supporting this research comprises structural responses of over 12,000 time history analyses for 100 SMRF buildings lumped into three categories: low-, mid-, and high-rise. The ground motions have been selected to represent three hazard levels, resulting in Service Level Earthquake (SLE), Design Basis Earthquake (DBE), and Maximum Considered Earthquake (MCE). Considering the primary period of each building and the target response spectra, a set of ground motions is selected, and the peak story drift ratios are extracted. Subsequently, unsupervised clustering techniques are employed to identify drift thresholds that distinguish between different damage states across various hazard levels, thereby refining the fixed boundaries recommended in existing codes and guidelines. Supervised learning techniques, on the other hand, are employed to predict the maximum drift ratio using features from ground motions and structural periods. The resulting drift ratio serves as an Engineering Demand Parameter (EDP), which, along with the hazard-informed drift threshold, is used to generate a machine learning-based fragility function. The proposed approach enables damage state identification of SMRF buildings under a specific ground motion, using only structural periods and signal features, without requiring detailed structural response data. The results of this study provide a set of site-specific hazard-based fragility curves, supporting seismic risk and loss assessment across different earthquake intensities.</div></div>","PeriodicalId":11763,"journal":{"name":"Engineering Structures","volume":"353 ","pages":"Article 122270"},"PeriodicalIF":6.4,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146116372","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-02-05DOI: 10.1016/j.engstruct.2026.122278
Sasa Cao , Xiaolong Sheng , Haojie Qiu , Osman E. Ozbulut
Conventional friction pendulum isolators rely on fixed spherical surfaces or discrete sliding stages, which constrain their ability to provide smooth stiffness adaptation and limit their frictional energy dissipation. To address these limitations, this study investigates a novel gravity-well double friction pendulum system (GW-DFPS) that employs a variable-curvature sliding surface to elongate displacement trajectories and enhance energy dissipation while enabling continuous stiffness softening at large displacements. Through a series of shake table experiments, a scaled bridge superstructure isolated with GW-DFPS was subjected to a range of uni- and bi-directional ground motions representing different site conditions and seismic intensities. Experimental results confirm that the system exhibits the intended softening behavior at larger displacements, effectively limiting force demands while accommodating significant lateral motions. Comparisons between unidirectional and bidirectional excitations highlight that the latter can lead to increased displacement demands, though with moderated acceleration responses. Residual displacements were small across all tests. Energy-based evaluations revealed a clear trade-off between kinetic and gravitational potential energy, with frictional dissipation increasing with sliding velocity. Overall, the GW-DFPS demonstrates strong potential as a next-generation seismic isolation device capable of sustaining large displacements while reducing shear forces transmitted to the superstructure.
{"title":"Shake table tests of gravity well-inspired double friction pendulum systems under Bi-directional ground motions","authors":"Sasa Cao , Xiaolong Sheng , Haojie Qiu , Osman E. Ozbulut","doi":"10.1016/j.engstruct.2026.122278","DOIUrl":"10.1016/j.engstruct.2026.122278","url":null,"abstract":"<div><div>Conventional friction pendulum isolators rely on fixed spherical surfaces or discrete sliding stages, which constrain their ability to provide smooth stiffness adaptation and limit their frictional energy dissipation. To address these limitations, this study investigates a novel gravity-well double friction pendulum system (GW-DFPS) that employs a variable-curvature sliding surface to elongate displacement trajectories and enhance energy dissipation while enabling continuous stiffness softening at large displacements. Through a series of shake table experiments, a scaled bridge superstructure isolated with GW-DFPS was subjected to a range of uni- and bi-directional ground motions representing different site conditions and seismic intensities. Experimental results confirm that the system exhibits the intended softening behavior at larger displacements, effectively limiting force demands while accommodating significant lateral motions. Comparisons between unidirectional and bidirectional excitations highlight that the latter can lead to increased displacement demands, though with moderated acceleration responses. Residual displacements were small across all tests. Energy-based evaluations revealed a clear trade-off between kinetic and gravitational potential energy, with frictional dissipation increasing with sliding velocity. Overall, the GW-DFPS demonstrates strong potential as a next-generation seismic isolation device capable of sustaining large displacements while reducing shear forces transmitted to the superstructure.</div></div>","PeriodicalId":11763,"journal":{"name":"Engineering Structures","volume":"353 ","pages":"Article 122278"},"PeriodicalIF":6.4,"publicationDate":"2026-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146116371","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号