Pub Date : 2025-11-01Epub Date: 2025-07-02DOI: 10.1016/j.advengsoft.2025.103977
Zepeng Wen , Xiaoya Zhai , Hongmei Kang
Topology optimization (TO) is a mature design technique that provides a conceptual design with the desired structural performance. However, the raw TO results cannot be interactively fine-tuned for conceptual design or further post-processing. This paper proposes an automatic and effective framework that can convert TO results into editable CAD models. The marching cubes algorithm is utilized to obtain discrete boundary points, which are subsequently transformed into CAD format via the sparse curve fitting technique. The proposed framework ensures robust and automatic reconstruction of the TO results of macrostructures and microstructures, especially complex small-scale structures with tiny holes and a few extracted points. We developed a Rhino plug-in, facilitating designers to modify the structural layout and intuitively assess the performance of the editable structures. Furthermore, we address stress concentrations by editing the reconstructed CAD models to illustrate the application of the proposed method.
{"title":"From density to CAD: Automatic and robust CAD model generation of topology optimization results via sparse optimization","authors":"Zepeng Wen , Xiaoya Zhai , Hongmei Kang","doi":"10.1016/j.advengsoft.2025.103977","DOIUrl":"10.1016/j.advengsoft.2025.103977","url":null,"abstract":"<div><div>Topology optimization (TO) is a mature design technique that provides a conceptual design with the desired structural performance. However, the raw TO results cannot be interactively fine-tuned for conceptual design or further post-processing. This paper proposes an automatic and effective framework that can convert TO results into editable CAD models. The marching cubes algorithm is utilized to obtain discrete boundary points, which are subsequently transformed into CAD format via the sparse curve fitting technique. The proposed framework ensures robust and automatic reconstruction of the TO results of macrostructures and microstructures, especially complex small-scale structures with tiny holes and a few extracted points. We developed a Rhino plug-in, facilitating designers to modify the structural layout and intuitively assess the performance of the editable structures. Furthermore, we address stress concentrations by editing the reconstructed CAD models to illustrate the application of the proposed method.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"209 ","pages":"Article 103977"},"PeriodicalIF":4.0,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144534815","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In urban civil engineering projects, deep contiguous piled walls are crucial for support and stability, particularly in limited spaces. This study employs advanced soft computing techniques, integrating FELA simulations, XGBoost, and optimization algorithms (PSO, GWO, and WOA), to predict the limiting uniform normal pressure behind circular and I-shaped contiguous piled walls in cohesive soil. The key factors in the prediction model include the soil properties, pile wall geometry, and soil strength parameters such as the spacing-to-width ratio (S/B or S/D), adhesion factor (m), height-to-width ratio of the I-section (H/B), and friction angle (φ). The sensitivity analysis reveals that for circular-shaped piles, φ is the most influential parameter (R = 0.69, Ri = 47 %), followed by the m and S/D, with moderate impacts (R = 0.49 and 0.28, Ri = 34 % and 19 %, respectively). For I-shaped piles, the m has the highest effect (R = 0.52, Ri = 28 %) followed by the S/B (R = 0.49, Ri = 26 %), H/B (R = 0.45, Ri = 24 %) and φ (R = 0.41, Ri = 22 %). The predictive performance of the constructed model was assessed using several performance metrics, regression plot, residual error histogram and regression error characteristic (REC) curve. The hybrid XGBoost-WOA model is identified as the most effective for both circular and I-shaped piles based on various performance metrics (R2, RMSE, MAE, WMAPE, VAF, and PI) and error analyses. This approach aims to improve prediction accuracy and address the limitations of traditional methods in deep contiguous piled wall design.
在城市土木工程项目中，深层连续桩墙对于支撑和稳定至关重要，特别是在有限的空间中。本研究采用先进的软计算技术，集成FELA模拟、XGBoost和优化算法（PSO、GWO和WOA），预测了粘性土中圆形和i形连续桩墙后的极限均匀法向压力。预测模型的关键因素包括土体性质、桩壁几何形状和土体强度参数，如间距比（S/B或S/D）、黏附系数(m)、工字段高宽比（H/B）、摩擦角（φ）等。敏感性分析表明，对于圆形桩，φ是影响最大的参数（R = 0.69, Ri = 47%），其次是m和S/D，影响中等（R = 0.49和0.28，Ri分别= 34%和19%）。对于工字形桩，m的影响最大（R = 0.52, Ri = 28%），其次是S/B （R = 0.49, Ri = 26%）、H/B （R = 0.45, Ri = 24%）和φ (R = 0.41, Ri = 22%)。采用多个性能指标、回归图、残差直方图和回归误差特征（REC）曲线对构建模型的预测性能进行评估。基于各种性能指标（R2、RMSE、MAE、WMAPE、VAF和PI）和误差分析，XGBoost-WOA混合模型被认为对圆形和i形桩都是最有效的。该方法旨在提高预测精度，解决传统方法在深层连续桩墙设计中的局限性。
{"title":"Prediction of the limiting uniform normal pressure in deep contiguous piled walls using soft computing techniques","authors":"Divesh Ranjan Kumar , Warit Wipulanusat , Duy Tan Tran , Suraparb Keawsawasvong","doi":"10.1016/j.advengsoft.2025.103993","DOIUrl":"10.1016/j.advengsoft.2025.103993","url":null,"abstract":"<div><div>In urban civil engineering projects, deep contiguous piled walls are crucial for support and stability, particularly in limited spaces. This study employs advanced soft computing techniques, integrating FELA simulations, XGBoost, and optimization algorithms (PSO, GWO, and WOA), to predict the limiting uniform normal pressure behind circular and I-shaped contiguous piled walls in cohesive soil. The key factors in the prediction model include the soil properties, pile wall geometry, and soil strength parameters such as the spacing-to-width ratio (S/B or S/D), adhesion factor (m), height-to-width ratio of the I-section (H/B), and friction angle (<em>φ</em>). The sensitivity analysis reveals that for circular-shaped piles, <em>φ</em> is the most influential parameter (<em>R</em> = 0.69, R<sub>i</sub> = 47 %), followed by the m and S/D, with moderate impacts (<em>R</em> = 0.49 and 0.28, R<sub>i</sub> = 34 % and 19 %, respectively). For I-shaped piles, the m has the highest effect (<em>R</em> = 0.52, R<sub>i</sub> = 28 %) followed by the S/B (<em>R</em> = 0.49, R<sub>i</sub> = 26 %), H/B (<em>R</em> = 0.45, R<sub>i</sub> = 24 %) and <em>φ</em> (<em>R</em> = 0.41, R<sub>i</sub> = 22 %). The predictive performance of the constructed model was assessed using several performance metrics, regression plot, residual error histogram and regression error characteristic (REC) curve. The hybrid XGBoost-WOA model is identified as the most effective for both circular and I-shaped piles based on various performance metrics (<em>R<sup>2</sup></em>, RMSE, MAE, WMAPE, VAF, and PI) and error analyses. This approach aims to improve prediction accuracy and address the limitations of traditional methods in deep contiguous piled wall design.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"209 ","pages":"Article 103993"},"PeriodicalIF":4.0,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144580068","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To investigate the effects of solar radiation temperature on steel-concrete composite beams in long-span bridges, a temperature analysis method based on meteorological parameters is proposed. First, an accurate thermal analysis model was established based on the principles of heat conduction and finite element computation theory. The model’s accuracy was validated using measured data. Subsequently, the representative values of temperature differences for the composite beam were calculated based on the analysis of historical meteorological data. The calculation method integrated the finite element temperature field model with the generalized extreme value distribution function. Finally, the simulated solar radiation temperature field was applied to the overall finite element model of the long-span bridge. The comprehensive influence of solar radiation temperature effects on long-span bridges was then evaluated. The results indicate that the finite element analysis results are in good agreement with the measured values. The 50-year return period representative value (Td) for the maximum vertical positive temperature difference in the composite beam is 17.613 °C. The compressive stress induced in the concrete bridge deck of the composite beam by temperature loads during the operational phase exceeds twice that caused by lane loads. By analyzing seven different load effect combinations, the impact of the most unfavorable load effect combination on the bridge structure was determined. This provides an important reference for bridge design and assessment. This research offers a novel method for understanding and predicting the temperature response of long-span bridges. Additionally, it provides theoretical foundations and technical support for achieving the lightweight goals of bridge health monitoring.
{"title":"Solar-induced thermal effects on steel-concrete composite beams: A meteorological data-driven analysis for long-span bridges","authors":"Yonghao Chu , Yuping Zhang , Chuanxi Li , Junliang Xie , Jiaping Jiang , Xing Tang","doi":"10.1016/j.advengsoft.2025.103995","DOIUrl":"10.1016/j.advengsoft.2025.103995","url":null,"abstract":"<div><div>To investigate the effects of solar radiation temperature on steel-concrete composite beams in long-span bridges, a temperature analysis method based on meteorological parameters is proposed. First, an accurate thermal analysis model was established based on the principles of heat conduction and finite element computation theory. The model’s accuracy was validated using measured data. Subsequently, the representative values of temperature differences for the composite beam were calculated based on the analysis of historical meteorological data. The calculation method integrated the finite element temperature field model with the generalized extreme value distribution function. Finally, the simulated solar radiation temperature field was applied to the overall finite element model of the long-span bridge. The comprehensive influence of solar radiation temperature effects on long-span bridges was then evaluated. The results indicate that the finite element analysis results are in good agreement with the measured values. The 50-year return period representative value (<em>T</em><sub>d</sub>) for the maximum vertical positive temperature difference in the composite beam is 17.613 °C. The compressive stress induced in the concrete bridge deck of the composite beam by temperature loads during the operational phase exceeds twice that caused by lane loads. By analyzing seven different load effect combinations, the impact of the most unfavorable load effect combination on the bridge structure was determined. This provides an important reference for bridge design and assessment. This research offers a novel method for understanding and predicting the temperature response of long-span bridges. Additionally, it provides theoretical foundations and technical support for achieving the lightweight goals of bridge health monitoring.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"209 ","pages":"Article 103995"},"PeriodicalIF":4.0,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711873","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2025-11-01Epub Date: 2025-08-05DOI: 10.1016/j.advengsoft.2025.103998
Xinming Li, Bowen Ji, Zhengdong Huang, Kuan Fan, Yuechen Hu, Jiachen Luo
The computational efficiency of stiffness matrix is commonly recognized as one of the primary challenges in mechanical analysis and optimization problems. In this paper, a tensorized framework is proposed to enhance the efficiency of stiffness matrix evaluations. The approach is validated through its application to isogeometric buckling optimization of laminated composite shells. Specifically, a matrix-oriented tensor multiplication (MOTM) is employed to facilitate parallel computation. Tensorized formulations for both stiffness matrix computation and sensitivity analysis are derived. Moreover, a comprehensive complexity analysis comparing the tensorized algorithm with conventional sequential algorithm is presented. Numerical examples illustrate that the proposed tensorized approach achieves a one-order-of-magnitude improvement in efficiency for stiffness matrix evaluations and a two-order-of-magnitude enhancement for sensitivity computations. Furthermore, this paper examines the elastic bound of lamination parameters (LPs), which are related to the positive definiteness of the elastic matrix. An artificial neural network (ANN) is integrated into the optimization process to enforce the elastic bound, thereby significantly reducing the number of indefinite elastic matrices at quadrature points.
{"title":"Tensorized computational framework for stiffness matrix and its application to buckling optimization of multi-patch laminated shells via isogeometric analysis","authors":"Xinming Li, Bowen Ji, Zhengdong Huang, Kuan Fan, Yuechen Hu, Jiachen Luo","doi":"10.1016/j.advengsoft.2025.103998","DOIUrl":"10.1016/j.advengsoft.2025.103998","url":null,"abstract":"<div><div>The computational efficiency of stiffness matrix is commonly recognized as one of the primary challenges in mechanical analysis and optimization problems. In this paper, a tensorized framework is proposed to enhance the efficiency of stiffness matrix evaluations. The approach is validated through its application to isogeometric buckling optimization of laminated composite shells. Specifically, a matrix-oriented tensor multiplication (MOTM) is employed to facilitate parallel computation. Tensorized formulations for both stiffness matrix computation and sensitivity analysis are derived. Moreover, a comprehensive complexity analysis comparing the tensorized algorithm with conventional sequential algorithm is presented. Numerical examples illustrate that the proposed tensorized approach achieves a one-order-of-magnitude improvement in efficiency for stiffness matrix evaluations and a two-order-of-magnitude enhancement for sensitivity computations. Furthermore, this paper examines the elastic bound of lamination parameters (LPs), which are related to the positive definiteness of the elastic matrix. An artificial neural network (ANN) is integrated into the optimization process to enforce the elastic bound, thereby significantly reducing the number of indefinite elastic matrices at quadrature points.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"209 ","pages":"Article 103998"},"PeriodicalIF":5.7,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144770595","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2025-11-01Epub Date: 2025-08-05DOI: 10.1016/j.advengsoft.2025.103997
Bangde Liu , Su Tian , Xin Liu , Frank Leone , Wenbin Yu
Tow-steered composites with curvilinear fiber paths offer enhanced mechanical performance in lightweight structures. However, the absence of commercial design tools for tow-steered composite structures limits innovation in their design for real-world applications. This paper introduces a user-friendly Design tool for Advanced Tailorable Composites (DATC), providing an integrated design framework in two widely used commercial finite element (FE) software packages, Abaqus and MSC.Patran/Nastran. DATC offers a graphical user interface (GUI) to connect multiscale plate modeling, FE modeling, machine learning (ML) modeling, and design optimization. The GUI streamlines the design process by managing all design configurations and interacting with several external codes. The multiscale modeling code SwiftComp calculates effective plate stiffness based on the steering fiber angles. The ML module trains efficient surrogate models as an alternative to FE models to reduce computational costs. The optimization module employs the open-source code Dakota to iteratively perform FE analysis with updated design variables, multiscale plate modeling, and optimization. The paper demonstrates the user-friendliness and adaptability of DATC through three case studies of tow-steered composite structures.
{"title":"Design tool for tow-steered composite laminates in Abaqus and MSC.Patran/Nastran","authors":"Bangde Liu , Su Tian , Xin Liu , Frank Leone , Wenbin Yu","doi":"10.1016/j.advengsoft.2025.103997","DOIUrl":"10.1016/j.advengsoft.2025.103997","url":null,"abstract":"<div><div>Tow-steered composites with curvilinear fiber paths offer enhanced mechanical performance in lightweight structures. However, the absence of commercial design tools for tow-steered composite structures limits innovation in their design for real-world applications. This paper introduces a user-friendly Design tool for Advanced Tailorable Composites (DATC), providing an integrated design framework in two widely used commercial finite element (FE) software packages, Abaqus and MSC.Patran/Nastran. DATC offers a graphical user interface (GUI) to connect multiscale plate modeling, FE modeling, machine learning (ML) modeling, and design optimization. The GUI streamlines the design process by managing all design configurations and interacting with several external codes. The multiscale modeling code SwiftComp calculates effective plate stiffness based on the steering fiber angles. The ML module trains efficient surrogate models as an alternative to FE models to reduce computational costs. The optimization module employs the open-source code Dakota to iteratively perform FE analysis with updated design variables, multiscale plate modeling, and optimization. The paper demonstrates the user-friendliness and adaptability of DATC through three case studies of tow-steered composite structures.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"209 ","pages":"Article 103997"},"PeriodicalIF":5.7,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144770597","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2025-11-01Epub Date: 2025-07-08DOI: 10.1016/j.advengsoft.2025.103982
Haochen Hu , Rui Zhong , Qingshan Wang
Based on dynamic vibration absorber (DVA) technology, a hybrid numerical method combining the spectro-geometric method (SGM) and the pseudo-excitation method (PEM) is proposed to analyze the free vibration characteristics and stochastic dynamic behavior of functionally graded graphene platelet-reinforced composites (FG-GPLRC) annular plates under various stationary and nonstationary stochastic excitations. The dynamic model of the DVA is simplified as a spring–mass system. Within the theoretical framework of the first-order shear deformation theory (FSDT), the Lagrangian energy functional of the coupled structure is constructed, and the dynamic response of the coupled system is obtained using the Rayleigh-Ritz variational method. The numerical accuracy of the proposed method is validated by comparing the results with those from existing literature and the finite element method (FEM). On this basis, the effectiveness of DVAs in suppressing structural vibration is demonstrated, and the effects of structural dimensions, material properties, and DVA installation parameters on the dynamic behavior of the coupled system are examined.
{"title":"Effect of distributed DVAs on stochastic dynamic behaviors of FG-GPLRC annular plates","authors":"Haochen Hu , Rui Zhong , Qingshan Wang","doi":"10.1016/j.advengsoft.2025.103982","DOIUrl":"10.1016/j.advengsoft.2025.103982","url":null,"abstract":"<div><div>Based on dynamic vibration absorber (DVA) technology, a hybrid numerical method combining the spectro-geometric method (SGM) and the pseudo-excitation method (PEM) is proposed to analyze the free vibration characteristics and stochastic dynamic behavior of functionally graded graphene platelet-reinforced composites (FG-GPLRC) annular plates under various stationary and nonstationary stochastic excitations. The dynamic model of the DVA is simplified as a spring–mass system. Within the theoretical framework of the first-order shear deformation theory (FSDT), the Lagrangian energy functional of the coupled structure is constructed, and the dynamic response of the coupled system is obtained using the Rayleigh-Ritz variational method. The numerical accuracy of the proposed method is validated by comparing the results with those from existing literature and the finite element method (FEM). On this basis, the effectiveness of DVAs in suppressing structural vibration is demonstrated, and the effects of structural dimensions, material properties, and DVA installation parameters on the dynamic behavior of the coupled system are examined.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"209 ","pages":"Article 103982"},"PeriodicalIF":4.0,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144580067","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Although several methods have been proposed for generating as-scanned point clouds, i.e. point clouds incorporating various realistic artefacts that would appear if the corresponding real objects were digitized for real, most of them still fail to take into account the complex phenomena that occur in a real acquisition devices. This paper presents a new way of artificially generating point clouds by combining simulation and machine learning. Starting from the CAD model of the object to be virtually scanned and from a scan configuration, structured light simulation first allows reconstructing a preliminary 3D point cloud. Then, a coverage prediction network is used to predict the regions that would be acquired if a real acquisition was to be done. The prediction model has been trained from a large database of scan configurations and point clouds scanned for real. Finally, filtering and cropping are performed to fine-tune the generated point cloud. Experiments confirm that this method can generate point clouds very close to those that a real scanner would acquire, as shown by several metrics characterizing both local and global similarity. Such a virtual scanning technique enables the rapid generation of large quantities of realistic point clouds, especially when compared to the time-consuming and costly processes involved in using physical acquisition systems. This opens up new perspectives in terms of access to realistic point cloud databases, in particular for the development of various AI-based approaches.
{"title":"As-scanned point cloud generation using structured-light simulation and machine learning-based coverage prediction","authors":"Tingcheng Li , Ruding Lou , Arnaud Polette , Manon Jubert , Dominique Nozais , Jean-Philippe Pernot","doi":"10.1016/j.advengsoft.2025.103996","DOIUrl":"10.1016/j.advengsoft.2025.103996","url":null,"abstract":"<div><div>Although several methods have been proposed for generating as-scanned point clouds, i.e. point clouds incorporating various realistic artefacts that would appear if the corresponding real objects were digitized for real, most of them still fail to take into account the complex phenomena that occur in a real acquisition devices. This paper presents a new way of artificially generating point clouds by combining simulation and machine learning. Starting from the CAD model of the object to be virtually scanned and from a scan configuration, structured light simulation first allows reconstructing a preliminary 3D point cloud. Then, a coverage prediction network is used to predict the regions that would be acquired if a real acquisition was to be done. The prediction model has been trained from a large database of scan configurations and point clouds scanned for real. Finally, filtering and cropping are performed to fine-tune the generated point cloud. Experiments confirm that this method can generate point clouds very close to those that a real scanner would acquire, as shown by several metrics characterizing both local and global similarity. Such a virtual scanning technique enables the rapid generation of large quantities of realistic point clouds, especially when compared to the time-consuming and costly processes involved in using physical acquisition systems. This opens up new perspectives in terms of access to realistic point cloud databases, in particular for the development of various AI-based approaches.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"209 ","pages":"Article 103996"},"PeriodicalIF":5.7,"publicationDate":"2025-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144723369","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2025-10-01Epub Date: 2025-06-13DOI: 10.1016/j.advengsoft.2025.103976
Jianxi Ou , Jianqin Zhang , Haoyu Li , Bin Duan
Intelligent road damage detection is critical for ensuring traffic safety and extending the lifespan of roads. However, existing methods struggle to balance high accuracy and real-time performance in complex detection scenarios and resource-constrained environments. To address this issue, this study proposes a lightweight multi-scale feature fusion model based on an improved YOLOv10—GAS-YOLO. The model utilizes a novel lightweight architecture (GSF-ST) designed through a combination of feature generation, asymmetric convolution, and grouped channel shuffling optimization strategies, significantly reducing computational complexity and parameter count while enhancing both global and local feature representation. To improve multi-scale damage detection performance, GAS-YOLO incorporates an improved bidirectional feature pyramid network (BiFPN) and Swin Transformer module. A resolution halving and channel doubling strategy enhances the detection ability of small targets. Moreover, the WiOU loss function further optimizes bounding box regression accuracy, mitigating errors caused by sample imbalance. Channel pruning techniques are applied to achieve secondary lightweight compression of the model, resulting in significant resource savings. Through comparative experiments and ablation analysis with several advanced damage detection models, this study demonstrates a significant performance improvement of GAS-YOLO. Experimental results show that GAS-YOLO exhibits outstanding performance in multi-scale damage detection tasks, with 5.6 M parameters, 8.4GFLOPs of computational complexity, and a model size of only 5.8 MB. Compared to baseline models, detection accuracy improves by 10.8 %, computational complexity is reduced by 2.57 times, and parameter count is reduced by 1.29 times, with an average detection accuracy of 86.5 % and a single image processing time of 6.1 ms. Validation on both public datasets and self-constructed datasets further proves its real-time processing capability while maintaining high accuracy. The GAS-YOLO model proposed in this study not only provides a practical solution for road damage detection in resource-constrained environments but also offers new insights for intelligent management of intelligent transportation and urban infrastructure, with broad application prospects.
{"title":"An improved YOLOv10-based lightweight multi-scale feature fusion model for road defect detection and its applications","authors":"Jianxi Ou , Jianqin Zhang , Haoyu Li , Bin Duan","doi":"10.1016/j.advengsoft.2025.103976","DOIUrl":"10.1016/j.advengsoft.2025.103976","url":null,"abstract":"<div><div>Intelligent road damage detection is critical for ensuring traffic safety and extending the lifespan of roads. However, existing methods struggle to balance high accuracy and real-time performance in complex detection scenarios and resource-constrained environments. To address this issue, this study proposes a lightweight multi-scale feature fusion model based on an improved YOLOv10—GAS-YOLO. The model utilizes a novel lightweight architecture (GSF-ST) designed through a combination of feature generation, asymmetric convolution, and grouped channel shuffling optimization strategies, significantly reducing computational complexity and parameter count while enhancing both global and local feature representation. To improve multi-scale damage detection performance, GAS-YOLO incorporates an improved bidirectional feature pyramid network (BiFPN) and Swin Transformer module. A resolution halving and channel doubling strategy enhances the detection ability of small targets. Moreover, the WiOU loss function further optimizes bounding box regression accuracy, mitigating errors caused by sample imbalance. Channel pruning techniques are applied to achieve secondary lightweight compression of the model, resulting in significant resource savings. Through comparative experiments and ablation analysis with several advanced damage detection models, this study demonstrates a significant performance improvement of GAS-YOLO. Experimental results show that GAS-YOLO exhibits outstanding performance in multi-scale damage detection tasks, with 5.6 M parameters, 8.4GFLOPs of computational complexity, and a model size of only 5.8 MB. Compared to baseline models, detection accuracy improves by 10.8 %, computational complexity is reduced by 2.57 times, and parameter count is reduced by 1.29 times, with an average detection accuracy of 86.5 % and a single image processing time of 6.1 ms. Validation on both public datasets and self-constructed datasets further proves its real-time processing capability while maintaining high accuracy. The GAS-YOLO model proposed in this study not only provides a practical solution for road damage detection in resource-constrained environments but also offers new insights for intelligent management of intelligent transportation and urban infrastructure, with broad application prospects.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"208 ","pages":"Article 103976"},"PeriodicalIF":4.0,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144270709","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2025-10-01Epub Date: 2025-06-26DOI: 10.1016/j.advengsoft.2025.103975
Indrajeet Sahu
The challenge of finding parametric coordinates of bilinear interpolation of a point with respect to a quadrilateral in 2D or 3D frequently arises as a subproblem in various applications, e.g. finite element methods, computational geometry, and computer graphics. The accuracy and efficiency of inverse mapping in such cases are critical, as the accumulation of errors can significantly affect the quality of the overall solution to the broader problem. This mapping is nonlinear and typically solved with Newton’s iterative method, which is not only prone to convergence issues but also incurs high computational cost. This paper presents an analytical solution to this inverse mapping, along with a comprehensive geometric analysis covering all possible quadrilateral configurations. It describes the invertibility of all points and extends the discussion to 3D and concave quadrilaterals. The proposed algorithm is robust, free from failure due to convergence issues or oscillations in iterative methods, and achieves approximately higher computational speed compared to Newton’s method for quadrilaterals with non-parallel opposite edges. This enables an efficient calculation of shape functions or interpolation functions at all invertible spatial points. The high-accuracy, high-speed computational solution will be particularly advantageous in applications involving high spatial or temporal discretisation (i.e. fine mesh and small timesteps) where iterative methods will be computationally expensive. The analytical solution based algorithm is available as an open-source library at https://github.com/sahu-indrajeet/Bilinear-Inverse-Mapper.
{"title":"Bilinear-inverse-mapper: Analytical solution and algorithm for inverse mapping of bilinear interpolation of quadrilaterals","authors":"Indrajeet Sahu","doi":"10.1016/j.advengsoft.2025.103975","DOIUrl":"10.1016/j.advengsoft.2025.103975","url":null,"abstract":"<div><div>The challenge of finding parametric coordinates of bilinear interpolation of a point with respect to a quadrilateral in 2D or 3D frequently arises as a subproblem in various applications, e.g. finite element methods, computational geometry, and computer graphics. The accuracy and efficiency of inverse mapping in such cases are critical, as the accumulation of errors can significantly affect the quality of the overall solution to the broader problem. This mapping is nonlinear and typically solved with Newton’s iterative method, which is not only prone to convergence issues but also incurs high computational cost. This paper presents an analytical solution to this inverse mapping, along with a comprehensive geometric analysis covering all possible quadrilateral configurations. It describes the invertibility of all points and extends the discussion to 3D and concave quadrilaterals. The proposed algorithm is robust, free from failure due to convergence issues or oscillations in iterative methods, and achieves approximately <span><math><mrow><mn>2</mn><mo>.</mo><mn>4</mn><mo>×</mo></mrow></math></span> higher computational speed compared to Newton’s method for quadrilaterals with non-parallel opposite edges. This enables an efficient calculation of shape functions or interpolation functions at all invertible spatial points. The high-accuracy, high-speed computational solution will be particularly advantageous in applications involving high spatial or temporal discretisation (i.e. fine mesh and small timesteps) where iterative methods will be computationally expensive. The analytical solution based algorithm is available as an open-source library at <span><span>https://github.com/sahu-indrajeet/Bilinear-Inverse-Mapper</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"208 ","pages":"Article 103975"},"PeriodicalIF":4.0,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144481081","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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 : 2025-10-01Epub Date: 2025-06-07DOI: 10.1016/j.advengsoft.2025.103971
Peng Ge , Ou Yang , Jia He , Zhiyu Liu , Hao Chen
Globally, the technique of reinforcing concrete structures with bonded fiber-reinforced polymers (FRP) has become widely adopted. The integrity of the interface between concrete and FRP significantly influences the behavior of the reinforced structure. Consequently, precise prediction of the bond strength at the concrete and FRP interface is crucial for the logical design and assessment of structures that are repaired and reinforced using FRP. This paper utilizes two emerging metaheuristic algorithms, the Slime Mould Algorithm (SMA) and the Dung Beetle Optimization Algorithm (DBO), to improve the performance of machine learning (ML) techniques, including KNN, SVR, GBDT, and XGBoost. Optimizing the ML models with metaheuristic algorithms significantly enhanced the prediction accuracy compared to the non-optimized models. The SMA-GBDT performed better than other ML models, achieving an R² of 0.9492, an MAE of 1.5294, an MSE of 6.4159, an RMSE of 2.5329, and a MAPE of 8.6916, based on the testing dataset. Specifically, the SMA-GBDT model exhibited improvements of 5.83%, 39.04%, 50.75%, 29.82%, and 43.84% in R², MAE, MSE, RMSE, and MAPE, respectively, compared to the non-optimized GBDT. The predictions made by the SMA-GBDT model were higher precision than those provided by the current design codes and existing models.
{"title":"Metaheuristic algorithms-optimized machine learning models for FRP-concrete interfacial bond strength prediction","authors":"Peng Ge , Ou Yang , Jia He , Zhiyu Liu , Hao Chen","doi":"10.1016/j.advengsoft.2025.103971","DOIUrl":"10.1016/j.advengsoft.2025.103971","url":null,"abstract":"<div><div>Globally, the technique of reinforcing concrete structures with bonded fiber-reinforced polymers (FRP) has become widely adopted. The integrity of the interface between concrete and FRP significantly influences the behavior of the reinforced structure. Consequently, precise prediction of the bond strength at the concrete and FRP interface is crucial for the logical design and assessment of structures that are repaired and reinforced using FRP. This paper utilizes two emerging metaheuristic algorithms, the Slime Mould Algorithm (SMA) and the Dung Beetle Optimization Algorithm (DBO), to improve the performance of machine learning (ML) techniques, including KNN, SVR, GBDT, and XGBoost. Optimizing the ML models with metaheuristic algorithms significantly enhanced the prediction accuracy compared to the non-optimized models. The SMA-GBDT performed better than other ML models, achieving an <em>R</em>² of 0.9492, an MAE of 1.5294, an MSE of 6.4159, an RMSE of 2.5329, and a MAPE of 8.6916, based on the testing dataset. Specifically, the SMA-GBDT model exhibited improvements of 5.83%, 39.04%, 50.75%, 29.82%, and 43.84% in <em>R</em>², MAE, MSE, RMSE, and MAPE, respectively, compared to the non-optimized GBDT. The predictions made by the SMA-GBDT model were higher precision than those provided by the current design codes and existing models.</div></div>","PeriodicalId":50866,"journal":{"name":"Advances in Engineering Software","volume":"208 ","pages":"Article 103971"},"PeriodicalIF":4.0,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144229745","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"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