Visual inspection is crucial for the maintenance of built infrastructures, facilitating early detection and quantification of damage. Traditional manual methods, however, often require inspectors to access dangerous or inaccessible areas, posing significant safety risks and inefficiencies. In response to these challenges, this paper introduces a portable visual inspection device (VID) integrated with three laser distance meters and a high-resolution camera. The VID enhances the efficiency of visual inspection by incorporating methods that accurately estimate the camera's pose relative to the target surface and determine a scale factor for precise damage quantification. The proposed methods were validated through experimental validations, demonstrating their precision and effectiveness. In lab-scale validation, the angle estimation showed accuracy with less than 3 degrees of error, and the scale factor estimation method showed discrepancies of less than 1 mm, even when the observation angle exceeded 20 degrees. Subsequent field experiments confirmed the VID's capability to detect and measure microcracks as narrow as 0.1 mm. Furthermore, the device successfully quantified non-crack damage with an error margin of 1.84%, even at challenging angles exceeding 45 degrees.
{"title":"Development of a portable device for structural visual inspection","authors":"Jongbin Won, Minhyuk Song, Jongwoong Park","doi":"10.1111/mice.13399","DOIUrl":"10.1111/mice.13399","url":null,"abstract":"<p>Visual inspection is crucial for the maintenance of built infrastructures, facilitating early detection and quantification of damage. Traditional manual methods, however, often require inspectors to access dangerous or inaccessible areas, posing significant safety risks and inefficiencies. In response to these challenges, this paper introduces a portable visual inspection device (VID) integrated with three laser distance meters and a high-resolution camera. The VID enhances the efficiency of visual inspection by incorporating methods that accurately estimate the camera's pose relative to the target surface and determine a scale factor for precise damage quantification. The proposed methods were validated through experimental validations, demonstrating their precision and effectiveness. In lab-scale validation, the angle estimation showed accuracy with less than 3 degrees of error, and the scale factor estimation method showed discrepancies of less than 1 mm, even when the observation angle exceeded 20 degrees. Subsequent field experiments confirmed the VID's capability to detect and measure microcracks as narrow as 0.1 mm. Furthermore, the device successfully quantified non-crack damage with an error margin of 1.84%, even at challenging angles exceeding 45 degrees.</p>","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"40 8","pages":"1061-1079"},"PeriodicalIF":8.5,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/mice.13399","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143532579","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Jianqi Zhang, Ling Ding, Wei Wang, Hainian Wang, Ioannis Brilakis, Diana Davletshina, Rauno Heikkilä, Xu Yang
Pavement crack measurement (PCM) is essential for automated, precise road condition assessment. However, balancing speed and accuracy on edge artificial intelligence (AI) mobile devices remains challenging. This paper proposes a real-time PCM framework for edge deployment, incorporating a lightweight distillation network and a surface feature measurement algorithm. Specifically, the proposed instance-aware hybrid distillation module combines feature-based and relation-based knowledge distillation, leveraging crack instance-related information for efficient knowledge transfer from teacher to student networks, which results in a more accurate and lightweight segmentation model. Additionally, a real-time crack surface feature measurement algorithm, based on distance mapping relationships and crack edge coordinate extraction, addresses issues with crack edge branching and loss, enhancing measurement efficiency. Real-time measurement was performed on actual roads utilizing mobile robot equipped with an edge computing unit. The crack segmentation precision reached 84.37%, with a frame per second of 77.72. Compared to the ground truth, the relative error for average crack width ranged from 6.42% to 40.65%, while the relative error for crack length varied between 1.48% and 3.76%. These findings highlight the feasibility of real-time crack assessment and save road maintenance costs.
{"title":"Crack segmentation-guided measurement with lightweight distillation network on edge device","authors":"Jianqi Zhang, Ling Ding, Wei Wang, Hainian Wang, Ioannis Brilakis, Diana Davletshina, Rauno Heikkilä, Xu Yang","doi":"10.1111/mice.13446","DOIUrl":"https://doi.org/10.1111/mice.13446","url":null,"abstract":"Pavement crack measurement (PCM) is essential for automated, precise road condition assessment. However, balancing speed and accuracy on edge artificial intelligence (AI) mobile devices remains challenging. This paper proposes a real-time PCM framework for edge deployment, incorporating a lightweight distillation network and a surface feature measurement algorithm. Specifically, the proposed instance-aware hybrid distillation module combines feature-based and relation-based knowledge distillation, leveraging crack instance-related information for efficient knowledge transfer from teacher to student networks, which results in a more accurate and lightweight segmentation model. Additionally, a real-time crack surface feature measurement algorithm, based on distance mapping relationships and crack edge coordinate extraction, addresses issues with crack edge branching and loss, enhancing measurement efficiency. Real-time measurement was performed on actual roads utilizing mobile robot equipped with an edge computing unit. The crack segmentation precision reached 84.37%, with a frame per second of 77.72. Compared to the ground truth, the relative error for average crack width ranged from 6.42% to 40.65%, while the relative error for crack length varied between 1.48% and 3.76%. These findings highlight the feasibility of real-time crack assessment and save road maintenance costs.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"46 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143539331","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}
Precision segmentation of cracks is important in industrial non-destructive testing, but the presence of shadows in the actual environment can interfere with the segmentation results of cracks. To solve this problem, this study proposes a two-stage domain adaptation framework called GAN-DANet for crack segmentation in shadowed environments. In the first stage, CrackGAN uses adversarial learning to merge features from shadow-free and shadowed datasets, creating a new dataset with more domain-invariant features. In the second stage, the CrackSeg network innovatively integrates enhanced Laplacian filtering (ELF) into high-resolution net to enhance crack edges and texture features while filtering out shadow information. In this model, CrackGAN addresses domain shift by generating a new dataset with domain-invariant features, avoiding direct feature alignment between source and target domains. The ELF module in CrackSeg effectively enhances crack features and suppresses shadow interference, improving the segmentation model's robustness in shadowed environments. Experiments show that GAN-DANet improves the crack segmentation accuracy, with the mean intersection over union value increasing from 57.87 to 75.03, which surpasses the performance of existing state-of-the-art domain adaptation algorithms.
{"title":"Generative adversarial network based on domain adaptation for crack segmentation in shadow environments","authors":"Yingchao Zhang, Cheng Liu","doi":"10.1111/mice.13451","DOIUrl":"https://doi.org/10.1111/mice.13451","url":null,"abstract":"Precision segmentation of cracks is important in industrial non-destructive testing, but the presence of shadows in the actual environment can interfere with the segmentation results of cracks. To solve this problem, this study proposes a two-stage domain adaptation framework called GAN-DANet for crack segmentation in shadowed environments. In the first stage, CrackGAN uses adversarial learning to merge features from shadow-free and shadowed datasets, creating a new dataset with more domain-invariant features. In the second stage, the CrackSeg network innovatively integrates enhanced Laplacian filtering (ELF) into high-resolution net to enhance crack edges and texture features while filtering out shadow information. In this model, CrackGAN addresses domain shift by generating a new dataset with domain-invariant features, avoiding direct feature alignment between source and target domains. The ELF module in CrackSeg effectively enhances crack features and suppresses shadow interference, improving the segmentation model's robustness in shadowed environments. Experiments show that GAN-DANet improves the crack segmentation accuracy, with the mean intersection over union value increasing from 57.87 to 75.03, which surpasses the performance of existing state-of-the-art domain adaptation algorithms.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"2 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-03-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143532414","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}
The cover image is based on the article Multifidelity graph neural networks for efficient and accurate mesh-based partial differential equations surrogate modeling by Negin Alemazkoor et al., https://doi.org/10.1111/mice.13312.
{"title":"Cover Image, Volume 40, Issue 7","authors":"","doi":"10.1111/mice.13449","DOIUrl":"https://doi.org/10.1111/mice.13449","url":null,"abstract":"<p><b>The cover image</b> is based on the article <i>Multifidelity graph neural networks for efficient and accurate mesh-based partial differential equations surrogate modeling</i> by Negin Alemazkoor et al., https://doi.org/10.1111/mice.13312.\u0000\u0000 <figure>\u0000 <div><picture>\u0000 <source></source></picture><p></p>\u0000 </div>\u0000 </figure>\u0000 </p>","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"40 7","pages":""},"PeriodicalIF":8.5,"publicationDate":"2025-03-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1111/mice.13449","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143527737","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Multivariate engineering formulas are the foundation of various engineering standards worldwide for constructing complex systems. Traditional formula discovery methods suffer from low efficiency, the curse of dimensionality, and low physical interpretability. To address these limitations, this study proposes a knowledge-based method for efficiently generating multivariate engineering formulas directly from data. The method consists of four components: (1) a deep generative model considering dimensional homogeneity, (2) a physics-adaptive normalization method for multiple engineering variables with different units, (3) a feature merging algorithm grounded in dimensionality theory, and (4) a machine learning-based data segmentation method for piecewise formulas. Experiments on two ground-truth datasets demonstrate that our proposed method improves the accuracy of the generated formulas by 35.6% (measured by mean absolute error), compared to the Eureqa program. Additionally, it enhances the mechanistic interpretability of the results, compared to both Eureqa and the emerging physics-informed neural network-based equation discovery methods. The piecewise formulas successfully capture the implicit mechanisms in the experimental data, consistent with theoretical analysis. Overall, our knowledge-based method holds great promise for improving the efficiency of discovering interpretable and generalizable multivariate engineering formulas, facilitating the transformation of new techniques from testing to applications.
{"title":"Multivariate engineering formulas discovery with knowledge-based neural network","authors":"Pei-Yao Chen, Chen Wang, Jian-Sheng Fan","doi":"10.1111/mice.13448","DOIUrl":"https://doi.org/10.1111/mice.13448","url":null,"abstract":"Multivariate engineering formulas are the foundation of various engineering standards worldwide for constructing complex systems. Traditional formula discovery methods suffer from low efficiency, the curse of dimensionality, and low physical interpretability. To address these limitations, this study proposes a knowledge-based method for efficiently generating multivariate engineering formulas directly from data. The method consists of four components: (1) a deep generative model considering dimensional homogeneity, (2) a physics-adaptive normalization method for multiple engineering variables with different units, (3) a feature merging algorithm grounded in dimensionality theory, and (4) a machine learning-based data segmentation method for piecewise formulas. Experiments on two ground-truth datasets demonstrate that our proposed method improves the accuracy of the generated formulas by 35.6% (measured by mean absolute error), compared to the <i>Eureqa</i> program. Additionally, it enhances the mechanistic interpretability of the results, compared to both <i>Eureqa</i> and the emerging physics-informed neural network-based equation discovery methods. The piecewise formulas successfully capture the implicit mechanisms in the experimental data, consistent with theoretical analysis. Overall, our knowledge-based method holds great promise for improving the efficiency of discovering interpretable and generalizable multivariate engineering formulas, facilitating the transformation of new techniques from testing to applications.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"210 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143495838","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}
Tao Wu, Shitong Hou, Zhishen Wu, Wen Xiong, Jian Zhang, Xinxing Shao, Xiaoyuan He, Gang Wu
Underwater structural inspection is essential for ensuring the safety and longevity of bridges. To improve the efficiency and accuracy of these inspections, this paper presents a method for measuring the morphology of bridge piers through refraction correction and multi-camera calibration. Using an underwater visual inspection platform with appropriate lighting, the measurement equipment mitigates low visibility challenges. A coplanar camera refraction parameter calibration method based on encoded markers is proposed to reduce the effects of refraction, along with the development of a multi-refraction correction model. Additionally, a novel multi-camera extrinsic calibration method is introduced to stitch point clouds. A comparative analysis of the two extrinsic calibration methods, conducted both in air and underwater, has been performed to validate the accuracy and efficiency of the proposed approach. Finally, the circular cross-section shape of the underwater bridge pier was successfully measured, and the results of defect localization were effectively presented.
{"title":"Underwater bridge pier morphology measurement method via refraction correction and multi-camera calibration","authors":"Tao Wu, Shitong Hou, Zhishen Wu, Wen Xiong, Jian Zhang, Xinxing Shao, Xiaoyuan He, Gang Wu","doi":"10.1111/mice.13440","DOIUrl":"https://doi.org/10.1111/mice.13440","url":null,"abstract":"Underwater structural inspection is essential for ensuring the safety and longevity of bridges. To improve the efficiency and accuracy of these inspections, this paper presents a method for measuring the morphology of bridge piers through refraction correction and multi-camera calibration. Using an underwater visual inspection platform with appropriate lighting, the measurement equipment mitigates low visibility challenges. A coplanar camera refraction parameter calibration method based on encoded markers is proposed to reduce the effects of refraction, along with the development of a multi-refraction correction model. Additionally, a novel multi-camera extrinsic calibration method is introduced to stitch point clouds. A comparative analysis of the two extrinsic calibration methods, conducted both in air and underwater, has been performed to validate the accuracy and efficiency of the proposed approach. Finally, the circular cross-section shape of the underwater bridge pier was successfully measured, and the results of defect localization were effectively presented.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"50 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143470502","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}
Abdullah M. Braik, Himadri Sen Gupta, Maria Koliou, Andrés D. González
Coastal communities are increasingly vulnerable to hurricanes, which cause billions of dollars in damage annually through wind, storm surge, and flooding. Mitigation efforts are essential to reduce these impacts but face significant challenges, including uncertainties in hazard prediction, damage estimation, and recovery costs. Resource constraints and the disproportionate burden borne by socioeconomically vulnerable groups further complicate retrofitting strategies. This study presents a probabilistic methodology to assess and mitigate hurricane risks by integrating hazard analysis, building fragility, and economic loss assessment. The methodology prioritizes retrofitting strategies using a risk-informed, equity-focused approach. Multi-objective optimization balances cost-effectiveness and risk reduction while promoting fair resource allocation among socioeconomic groups. The novelty of this study lies in its direct integration of equity as an objective in resource allocation through multi-objective optimization, its comprehensive consideration of multi-hazard risks, its inclusion of both direct and indirect losses in cost assessments, and its use of probabilistic hazard analysis to incorporate varying time horizons. A case study of the Galveston testbed demonstrates the methodology's potential to minimize damage and foster equitable resilience. Analysis of budget scenarios and trade-offs between cost and equity underscores the importance of comprehensive loss assessments and equity considerations in mitigation and resilience planning. Key findings highlight the varied effectiveness of retrofitting strategies across different budgets and time horizons, the necessity of addressing both direct and indirect losses, and the importance of multi-hazard considerations for accurate risk assessments. Multi-objective optimization underscores that equitable solutions are achievable even under constrained budgets. Beyond a certain point, achieving equity does not necessarily increase expected losses, demonstrating that more equitable solutions can be implemented without compromising overall cost-effectiveness.
{"title":"Multi-hazard probabilistic risk assessment and equitable multi-objective optimization of building retrofit strategies in hurricane-vulnerable communities","authors":"Abdullah M. Braik, Himadri Sen Gupta, Maria Koliou, Andrés D. González","doi":"10.1111/mice.13445","DOIUrl":"https://doi.org/10.1111/mice.13445","url":null,"abstract":"Coastal communities are increasingly vulnerable to hurricanes, which cause billions of dollars in damage annually through wind, storm surge, and flooding. Mitigation efforts are essential to reduce these impacts but face significant challenges, including uncertainties in hazard prediction, damage estimation, and recovery costs. Resource constraints and the disproportionate burden borne by socioeconomically vulnerable groups further complicate retrofitting strategies. This study presents a probabilistic methodology to assess and mitigate hurricane risks by integrating hazard analysis, building fragility, and economic loss assessment. The methodology prioritizes retrofitting strategies using a risk-informed, equity-focused approach. Multi-objective optimization balances cost-effectiveness and risk reduction while promoting fair resource allocation among socioeconomic groups. The novelty of this study lies in its direct integration of equity as an objective in resource allocation through multi-objective optimization, its comprehensive consideration of multi-hazard risks, its inclusion of both direct and indirect losses in cost assessments, and its use of probabilistic hazard analysis to incorporate varying time horizons. A case study of the Galveston testbed demonstrates the methodology's potential to minimize damage and foster equitable resilience. Analysis of budget scenarios and trade-offs between cost and equity underscores the importance of comprehensive loss assessments and equity considerations in mitigation and resilience planning. Key findings highlight the varied effectiveness of retrofitting strategies across different budgets and time horizons, the necessity of addressing both direct and indirect losses, and the importance of multi-hazard considerations for accurate risk assessments. Multi-objective optimization underscores that equitable solutions are achievable even under constrained budgets. Beyond a certain point, achieving equity does not necessarily increase expected losses, demonstrating that more equitable solutions can be implemented without compromising overall cost-effectiveness.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"12 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143463166","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}
Cracking impacts asphalt concrete durability primarily due to cohesive asphalt binder failures. The poker chip test has recently been introduced to better characterize the cracking potential of asphalt binders by fracturing a specimen in a realistic stress state to a thin binder film. However, broader adoption faces challenges due to high instrumentation costs for measuring load and displacement. This paper presents and validates a deep learning model that predicts ductility and tensile strength from posttest images of fractured binder surfaces, with potential extensions to simplified instrumentation. The hybrid model, named PCNet, integrates a custom lightweight convolutional neural network (CNN) developed to capture local features (e.g., edges, boundaries, contours) within fracture cavities with a Swin Transformer that models global contextual dependencies. A bidirectional cross-attention fusion module is designed to facilitate mutual information exchange between CNN and transformer branches. The fused features are then processed by a fully connected network (FCN) to predict indices derived from the test. The proposed model demonstrates high predictive accuracy across a range of binders and test configurations, achieving an <span data-altimg="/cms/asset/e9fbf493-59db-40fa-93fd-5f973187445b/mice13447-math-0001.png"></span><mjx-container ctxtmenu_counter="150" ctxtmenu_oldtabindex="1" jax="CHTML" role="application" sre-explorer- style="font-size: 103%; position: relative;" tabindex="0"><mjx-math aria-hidden="true" location="graphic/mice13447-math-0001.png"><mjx-semantics><mjx-msup data-semantic-children="0,1" data-semantic- data-semantic-role="latinletter" data-semantic-speech="upper R squared" data-semantic-type="superscript"><mjx-mi data-semantic-annotation="clearspeak:simple" data-semantic-font="italic" data-semantic- data-semantic-parent="2" data-semantic-role="latinletter" data-semantic-type="identifier"><mjx-c></mjx-c></mjx-mi><mjx-script style="vertical-align: 0.363em;"><mjx-mn data-semantic-annotation="clearspeak:simple" data-semantic-font="normal" data-semantic- data-semantic-parent="2" data-semantic-role="integer" data-semantic-type="number" size="s"><mjx-c></mjx-c></mjx-mn></mjx-script></mjx-msup></mjx-semantics></mjx-math><mjx-assistive-mml display="inline" unselectable="on"><math altimg="urn:x-wiley:10939687:media:mice13447:mice13447-math-0001" display="inline" location="graphic/mice13447-math-0001.png" xmlns="http://www.w3.org/1998/Math/MathML"><semantics><msup data-semantic-="" data-semantic-children="0,1" data-semantic-role="latinletter" data-semantic-speech="upper R squared" data-semantic-type="superscript"><mi data-semantic-="" data-semantic-annotation="clearspeak:simple" data-semantic-font="italic" data-semantic-parent="2" data-semantic-role="latinletter" data-semantic-type="identifier">R</mi><mn data-semantic-="" data-semantic-annotation="clearspeak:simple" data-semantic-font="normal" data-semantic-parent="2" data-semantic-rol
{"title":"Hybrid deep learning model for predicting failure properties of asphalt binder from fracture surface images","authors":"Babak Asadi, Viraj Shah, Abhilash Vyas, Mani Golparvar-Fard, Ramez Hajj","doi":"10.1111/mice.13447","DOIUrl":"https://doi.org/10.1111/mice.13447","url":null,"abstract":"Cracking impacts asphalt concrete durability primarily due to cohesive asphalt binder failures. The poker chip test has recently been introduced to better characterize the cracking potential of asphalt binders by fracturing a specimen in a realistic stress state to a thin binder film. However, broader adoption faces challenges due to high instrumentation costs for measuring load and displacement. This paper presents and validates a deep learning model that predicts ductility and tensile strength from posttest images of fractured binder surfaces, with potential extensions to simplified instrumentation. The hybrid model, named PCNet, integrates a custom lightweight convolutional neural network (CNN) developed to capture local features (e.g., edges, boundaries, contours) within fracture cavities with a Swin Transformer that models global contextual dependencies. A bidirectional cross-attention fusion module is designed to facilitate mutual information exchange between CNN and transformer branches. The fused features are then processed by a fully connected network (FCN) to predict indices derived from the test. The proposed model demonstrates high predictive accuracy across a range of binders and test configurations, achieving an <span data-altimg=\"/cms/asset/e9fbf493-59db-40fa-93fd-5f973187445b/mice13447-math-0001.png\"></span><mjx-container ctxtmenu_counter=\"150\" ctxtmenu_oldtabindex=\"1\" jax=\"CHTML\" role=\"application\" sre-explorer- style=\"font-size: 103%; position: relative;\" tabindex=\"0\"><mjx-math aria-hidden=\"true\" location=\"graphic/mice13447-math-0001.png\"><mjx-semantics><mjx-msup data-semantic-children=\"0,1\" data-semantic- data-semantic-role=\"latinletter\" data-semantic-speech=\"upper R squared\" data-semantic-type=\"superscript\"><mjx-mi data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic- data-semantic-parent=\"2\" data-semantic-role=\"latinletter\" data-semantic-type=\"identifier\"><mjx-c></mjx-c></mjx-mi><mjx-script style=\"vertical-align: 0.363em;\"><mjx-mn data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"normal\" data-semantic- data-semantic-parent=\"2\" data-semantic-role=\"integer\" data-semantic-type=\"number\" size=\"s\"><mjx-c></mjx-c></mjx-mn></mjx-script></mjx-msup></mjx-semantics></mjx-math><mjx-assistive-mml display=\"inline\" unselectable=\"on\"><math altimg=\"urn:x-wiley:10939687:media:mice13447:mice13447-math-0001\" display=\"inline\" location=\"graphic/mice13447-math-0001.png\" xmlns=\"http://www.w3.org/1998/Math/MathML\"><semantics><msup data-semantic-=\"\" data-semantic-children=\"0,1\" data-semantic-role=\"latinletter\" data-semantic-speech=\"upper R squared\" data-semantic-type=\"superscript\"><mi data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic-parent=\"2\" data-semantic-role=\"latinletter\" data-semantic-type=\"identifier\">R</mi><mn data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"normal\" data-semantic-parent=\"2\" data-semantic-rol","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"22 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143463167","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}
Yanjia Wang, Dong Yang, Ye Yuan, Jing Zhang, Francis T. K. Au
This paper introduces a novel hybrid multi‐model thermo‐temporal physics‐informed neural network (TT‐PINN) framework for thermal loading prediction in composite bridge decks. Unlike the existing PINN applications in heat transfer that focus on simple geometries, this framework uniquely addresses multi‐material domains and realistic boundary conditions through a dual‐network architecture designed for composite structures. The framework further incorporates the environmental boundary conditions of natural convection and solar radiation into the loss function and employs transfer learning for efficient adaptation to varying conditions. Moreover, a transfer learning mechanism enables rapid adaptation to new environmental states, thus markedly reducing the computations as compared to the conventional finite element method (FEM). Through noise‐augmented training and parameter identification, the TT‐PINN effectively handles the real‐world monitoring data uncertainties and allows material property calibration with limited sensor data. The framework's ability to capture complex thermal behavior is validated by studying a cable‐stayed bridge. It significantly reduces the computational costs as compared to the traditional FEM approaches.
{"title":"Hybrid physics‐informed neural network with parametric identification for modeling bridge temperature distribution","authors":"Yanjia Wang, Dong Yang, Ye Yuan, Jing Zhang, Francis T. K. Au","doi":"10.1111/mice.13436","DOIUrl":"https://doi.org/10.1111/mice.13436","url":null,"abstract":"This paper introduces a novel hybrid multi‐model thermo‐temporal physics‐informed neural network (TT‐PINN) framework for thermal loading prediction in composite bridge decks. Unlike the existing PINN applications in heat transfer that focus on simple geometries, this framework uniquely addresses multi‐material domains and realistic boundary conditions through a dual‐network architecture designed for composite structures. The framework further incorporates the environmental boundary conditions of natural convection and solar radiation into the loss function and employs transfer learning for efficient adaptation to varying conditions. Moreover, a transfer learning mechanism enables rapid adaptation to new environmental states, thus markedly reducing the computations as compared to the conventional finite element method (FEM). Through noise‐augmented training and parameter identification, the TT‐PINN effectively handles the real‐world monitoring data uncertainties and allows material property calibration with limited sensor data. The framework's ability to capture complex thermal behavior is validated by studying a cable‐stayed bridge. It significantly reduces the computational costs as compared to the traditional FEM approaches.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"12 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143435073","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}
Road cracks pose a serious threat to the stability of road structures and traffic safety. Therefore, this paper proposes an optimized accurate road crack segmentation network called MBGBNet, which can solve the problems of complex background, tiny cracks, and irregular edges in road segmentation. First, multi-scale domain feature aggregation is proposed to address the interference of complex background. Second, bidirectional embedding fusion adaptive attention is proposed to capture the features of tiny cracks, and finally, Gaussian weighted edge segmentation algorithm is proposed to ensure the accuracy of crack edge segmentation. In addition, this paper uses the preheated bat optimization algorithm, which can quickly determine the optimal learning rate to converge the equilibrium. In the validation experiments on the self-built dataset, mean intersection over union reaches 80.54% and precision of 86.38%. MBGBNet outperforms the other seven state-of-the-art crack segmentation networks on the three classical crack datasets, highlighting its advanced segmentation capabilities. Therefore, MBGBNet is an effective auxiliary method for solving road safety problems.
{"title":"An optimized and precise road crack segmentation network in complex scenarios","authors":"Gang Wang, MingFang He, Genhua Liu, Liujun Li, Exian Liu, Guoxiong Zhou","doi":"10.1111/mice.13444","DOIUrl":"https://doi.org/10.1111/mice.13444","url":null,"abstract":"Road cracks pose a serious threat to the stability of road structures and traffic safety. Therefore, this paper proposes an optimized accurate road crack segmentation network called MBGBNet, which can solve the problems of complex background, tiny cracks, and irregular edges in road segmentation. First, multi-scale domain feature aggregation is proposed to address the interference of complex background. Second, bidirectional embedding fusion adaptive attention is proposed to capture the features of tiny cracks, and finally, Gaussian weighted edge segmentation algorithm is proposed to ensure the accuracy of crack edge segmentation. In addition, this paper uses the preheated bat optimization algorithm, which can quickly determine the optimal learning rate to converge the equilibrium. In the validation experiments on the self-built dataset, mean intersection over union reaches 80.54% and precision of 86.38%. MBGBNet outperforms the other seven state-of-the-art crack segmentation networks on the three classical crack datasets, highlighting its advanced segmentation capabilities. Therefore, MBGBNet is an effective auxiliary method for solving road safety problems.","PeriodicalId":156,"journal":{"name":"Computer-Aided Civil and Infrastructure Engineering","volume":"52 1","pages":""},"PeriodicalIF":11.775,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143427302","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