检索结果-内蒙古大学图书馆

11th International Conference on Bridge Maintenance, Safety and Management, IABMAS 2022

作者： Deng, L. Chu, H.H. Wang, W. Key Laboratory for Damage Diagnosis of Engineering Structures of Hunan Province Hunan University Changsha China

ISBN: (纸本)9781032345314

Recent studies have employed convolutional neural networks (CNNs) to boost detection efficiency of cracks in bridge inspection. Class imbalance causes poor segmentation results for tiny cracks. Moreover, the local receptive field limits the ability of CNNs to combine local and global dependencies, affecting the detection accuracy of tiny cracks throughout the entire image. To tackle these issues in tiny crack segmentation, an encoder-decoder architecture with feature fusion and attention mechanisms is proposed. The improved residual network is used to capture the local features of tiny cracks. The dual attention module is then incorporated into the architecture, allowing for a more precise separation of tiny cracks from the surroundings. Additionally, a multi-scale fusion process is proposed to preserve the details of tiny cracks' edges. Finally, a joint learning loss of the cross-entropy and similarity is proposed to ameliorate the poor convergence induced by the severe class imbalance of tiny crack images. This work significantly improves the capability for tiny crack segmentation compared with four typical networks on two open-source crack datasets, achieving a dice score of 87.88%, which is at least 7.55% higher than the four state-of-the-art (SOTA) CNN-based methods. © 2022 the Author(s).

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

A Simple Analytical Approach for Stress Calculation of Solar Cells 10

A Simple Analytical Approach for Stress Calculation of Solar...

引用

10th Annual International Conference on Material Science and Environmental engineering, MSEE 2022

作者： Li, Zhiqi Key Laboratory for Damage Diagnosis of Engineering Structures of Hunan Province College of Civil Engineering Hunan University Changsha410082 China

A simple analytical solution was proposed for stress evaluation of the solar cells. The present solution was then applied to the stress calculation of the cells in 1580mm×808mm PV module, the variation of the wind pressures and the shear storage modulus of the EVA were taken into account in calculation. The results from the present solution show in good agreement with those by FE. © Published under licence by IOP Publishing Ltd.

关键词： Solar panels

来源：评论

学校读者我要写书评

暂无评论

Resonances of highway bridges induced by autonomous truck platooning 11th

Resonances of highway bridges induced by autonomous truck pl...

引用

11th International Conference on Bridge Maintenance, Safety and Management, IABMAS 2022

作者： Ling, T.Y. He, W. Wu, H.B. Deng, L. College of Civil Engineering Hunan University Hunan Province Changsha China Key Laboratory for Damage Diagnosis of Engineering Structures of Hunan Province Hunan University Hunan Changsha China

ISBN: (纸本)9781032345314

Autonomous truck platooning (ATP) used in road transport helps achieve emission peak and carbon neutrality. For this reason, increasing research investments have been added to ATP recently so as to accelerate its commercial application. However, the unusual load patterns of ATP on highway bridges have not been explicitly considered in current bridge design methods. Although some evaluations regarding ATP-bridge have been preliminarily conducted, they were focused on the static effect. In this paper, the critical dynamic loading scenarios, particularly in resonance conditions, during ATP crossing highway bridge were investigated via both theoretical deduction and numerical simulation. Results showed that highway bridge resonances easily happen when ATP crosses the bridge at normal speeds. As a result, the impact factors (IMs) of the bridge can be higher than those caused by a single truck and also exceed the code-specified values, which may thus require an improvement. © 2022 the Author(s).

关键词： Trucks

来源：评论

学校读者我要写书评

暂无评论

A New Method for Calculating Bridge Influence Surface

A New Method for Calculating Bridge Influence Surface

引用

IABSE Congress Nanjing 2022 - Bridges and structures: Connection, Integration and Harmonisation

作者： Wu, Haibing He, Wei Deng, Lu College of Civil Engineering Hunan University Hunan Province Changsha China Key Laboratory for Damage Diagnosis of Engineering Structures of Hunan Province Hunan University Hunan Changsha China

ISBN: (纸本)9783857481840

Bridge influence surface (BIS), which reflects the relationship between the bridge response and the unit load moving on the bridge, has been commonly used in bridge weigh-in-motion, bridge damage detection, etc. Generally, the BIS can be obtained by interpolating bridge influence lines (BILs) at different lateral positions. In field tests, the BILs are usually extracted from the bridge response induced by a moving calibration vehicle with known axle loads. However, when taking the axle loads of the calibration vehicle into calculation, the impacts of the transverse distance between coaxial wheels and the unbalance of the coaxial wheel loads are ignored. Hence, errors may be brought to the calculated BIL and then propagated to the BIS. To remove these impacts, a new BIS calculation method, which takes the load of each wheel rather than each axle into calculation, is proposed in this research. Numerical simulation shows that the BIS can be successfully estimated by this method. © IABSE Congress Nanjing 2022 - Bridges and structures: Connection, Integration and Harmonisation, Report. All rights reserved.

关键词： Numerical methods

来源：评论

学校读者我要写书评

暂无评论

3D keypoint detection-based automated rebar spacing inspection: Application for robotic integration

引用

Advanced engineering Informatics 2025年 66卷

作者： Deng, Lu Wang, Songyue Guo, Jingjing Cao, Ran Liu, Mi Key Laboratory of Damage Diagnosis for Engineering Structures of Hunan Province Hunan University Changsha China College of Civil Engineering Hunan University Changsha China

Ensuring the rebar installation quality is crucial for the strength and durability of reinforced concrete structures. Existing inspection methods fail to meet the real-time and efficiency demands of contemporary robot-assisted construction. To bridge this gap, this paper proposes a rebar spacing inspection method based on three-dimensional (3D) keypoint detection. This method consists of three main modules: (1) target rebar layer extraction, which eliminates interference from background rebars and allows for flexible inspection without fixed shooting distance and angle;(2) rebar crosspoint recognition and localization (RCRL), which combines RCRL-YOLOv8 Pose keypoint detection with depth information for accurate 3D crosspoint coordinates;and (3) automatic rebar spacing inspection, which identifies all adjacent crosspoints and calculates spacing values. The integration of these three modules facilitates an end-to-end inspection process, from image input to result output, and is further embedded into robotic construction systems to achieve automated spacing inspection in future real-world applications. The effectiveness and robustness of the proposed method were validated on physical rebar frames, achieving a mean Average Precision (mAP) of 98.5 % in crosspoint localization and an average error in rebar spacing inspection of 1.26 mm. Additionally, applications for real-world scenarios and robotic integration substantiated the method's feasibility, underscoring its potential for practical industrial applications. © 2025 Elsevier Ltd

关键词： damage detection

来源：评论

学校读者我要写书评

暂无评论

Automated Crack Detection Method Based on Deep Learning and 3D Reconstruction for Concrete Bridges

Automated Crack Detection Method Based on Deep Learning and ...

引用

IABSE Congress Nanjing 2022 - Bridges and structures: Connection, Integration and Harmonisation

作者： Sun, Tao Deng, Lu Cao, Ran Wang, Wei College of Civil Engineering Hunan University Hunan Province Changsha China Key Laboratory for Damage Diagnosis of Engineering Structures of Hunan Province China Hunan University Changsha China

ISBN: (纸本)9783857481840

Automated image-based bridge crack detection, as a promising technique, can be used to overcome the limitations of human visual inspection. However, results from current image-based methods are generally localized and lack 3D geometric information, which makes it difficult for structural assessment. To solve this issue, a crack detection method that combines deep learning and 3D reconstruction is proposed in this paper. Firstly, a 2D feature-based approach is developed to extract keyframes from the video adaptively. Secondly, a segmentation network is implemented to conduct pixel-level crack segmentation. Finally, image-based 3D reconstruction and crack mapping are used to create the 3D structure model with crack semantics. A field experiment is also carried out on an in-service concrete bridge for validation and discussion of the proposed method. The 3D model created by the proposed method can significantly improve the crack inspection of concrete bridges. © IABSE Congress Nanjing 2022 - Bridges and structures: Connection, Integration and Harmonisation, Report. All rights reserved.

关键词： Crack detection

来源：评论

学校读者我要写书评

暂无评论

Fatigue Reliability Analysis and Traffic Load Control of Steel Bridges Based on Artificial Neural Network

Fatigue Reliability Analysis and Traffic Load Control of Ste...

引用

2022 International Conference on Smart Transportation and City engineering, STCE 2022

作者： Nie, Lei Wang, Wei Deng, Lu Hunan University Hunan Changsha410012 China Key Laboratory for Damage Diagnosis of Engineering Structures of Hunan Province Hunan University Hunan Changsha410012 China

ISBN: (纸本)9781510660250

Steel bridges have the advantages of light weight, high strength, prefabricated construction, and short construction time, and therefore have been widely used in many countries. In conventional bridge design, the load bearing capacity of the structure is considered as the most important safety factor. However, as the service life increases, the actual load-carrying capacity of bridges gradually decreases due to the combined action of the environmental corrosion and repeated vehicle loads, resulting in shortened bridge service life. In this paper, the fatigue reliability index of a steel girder bridge over its whole life is investigated based on artificial neural networks. The effects of truck traffic load and environmental corrosivity on the fatigue life of the steel girder bridge are analyzed and measures to control the traffic load are discussed. The research results can serve as a reference for traffic load management of highway steel bridges. © 2022 SPIE.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

Automatic Inspection of Cracks on Pavement Surfaces Based on Improved Segmentation Model

Automatic Inspection of Cracks on Pavement Surfaces Based on...

引用

2022 International Conference on Smart Transportation and City engineering, STCE 2022

作者： Deng, Lu Zhao, Xu Chu, Honghu Xiang, Chao Key Laboratory for Damage Diagnosis of Engineering Structures of Hunan Province Hunan University Hunan Changsha410012 China Hunan University Hunan Changsha410012 China

ISBN: (纸本)9781510660250

Periodic crack inspections are greatly required for the operation and maintenance of the pavement. With the rapid development of the deep learning (DL) method and the computer hardware, manual visual inspections have been gradually replaced by DL-based inspections due to their high efficiency and low cost. In this study, a novel segmentation model was proposed to conduct precise identification of cracks. Firstly, the YOLOv5 was adopted as basic detector for the coarse prediction of the crack type and crack location. Secondly, a mask branch was customized and incorporated into the aforementioned detector for parallel mask production. In addition, the Neck, a feature fusion component, was upgraded to produce higher-quality masks, and a sampling-free approach was investigated to reduce class imbalance's influence. Compared to a normal segmentation model, the proposed method demonstrated a better balance between accuracy and speed, with a speed of 33.2 frames per second and an accuracy of 87.13 mAP. © 2022 SPIE.

关键词： Crack detection

来源：评论

学校读者我要写书评

暂无评论

Experimental Study on Progressive Collapse Behavior of Frame structures Triggered by Impact Column Removal

SSRN

引用

SSRN 2024年

作者： Yi, Fan Yi, Wei-Jian Sun, Jing-Ming Zhou, Yun Zhang, Wang-Xi He, Qing-Feng College of Civil Engineering Hunan Provincial Key Lab on Damage Diagnosis for Engineering Structures Hunan University Changsha410082 China

In the research field investigating the progressive collapse of building structures, event-dependent collapse processes have gained increasing attention in analytical and numerical studies. Different events could cause varying effects on progressive collapse resistance. The alternate load path could be related to events, and the event itself could also lead to variations in load actions. However, experimental studies on reinforced concrete (RC) frame structures triggering structural column removal by the low-velocity impact have not been reported. Therefore, this paper performed an initial experimental study employing impact loading as the extreme event to investigate subsequent progressive collapse behavior of structures. Tests were conducted on six RC substructures consisting of a two-span beam and a structural column. Gravity loads were applied to the top of substructures, and a pendulum impact setup was utilized to remove RC columns by low-velocity impact. When the column underwent lateral failures, the downward force exerted by longitudinal steel bars of the column, before they fractured, pulled the two-span beam beyond the compressive arch action (CAA) stage, leading to a collapse process entirely different from their event-independent counterparts. The parametric study based on experimental results indicated that low-elevation impact and increase in the cross-sectional areas of column longitudinal bars detrimentally affected the performance of two-span beams resisting progressive collapse, while the increase in concrete strength of specimens partially improved the residual bearing capacity after impact column removal (ICR). Based on a dynamic model, a simplified calculation method is proposed for quantifying the downward force. © 2024, The Authors. All rights reserved.

关键词： Reinforced concrete

来源：评论

学校读者我要写书评

暂无评论

Response probability distribution estimation of expensive computer simulators: A Bayesian active learning perspective using Gaussian process regression

arXiv

引用

arXiv 2024年

作者： Dang, Chao Valdebenito, Marcos A. Manque, Nataly A. Xu, Jun Faes, Matthias G.R. TU Dortmund University Leonhard-Euler-Str. 5 Dortmund44227 Germany College of Civil Engineering Hunan University Changsha410082 China Key Lab on Damage Diagnosis for Engineering Structures of Hunan Province Changsha410082 China

Estimation of the response probability distributions of computer simulators in the presence of randomness is a crucial task in many fields. However, achieving this task with guaranteed accuracy remains an open computational challenge, especially for expensive-to-evaluate computer simulators. In this work, a Bayesian active learning perspective is presented to address the challenge, which is based on the use of the Gaussian process (GP) regression. First, estimation of the response probability distributions is conceptually interpreted as a Bayesian inference problem, as opposed to frequentist inference. This interpretation provides several important benefits: (1) it quantifies and propagates discretization error probabilistically;(2) it incorporates prior knowledge of the computer simulator, and (3) it enables the effective reduction of numerical uncertainty in the solution to a prescribed level. The conceptual Bayesian idea is then realized by using the GP regression, where we derive the posterior statistics of the response probability distributions in semi-analytical form and also provide a numerical solution scheme. Based on the practical Bayesian approach, a Bayesian active learning (BAL) method is further proposed for estimating the response probability distributions. In this context, the key contribution lies in the development of two crucial components for active learning, i.e., stopping criterion and learning function, by taking advantage of posterior statistics. It is empirically demonstrated by five numerical examples that the proposed BAL method can efficiently estimate the response probability distributions with desired accuracy. © 2024, CC BY.

关键词： Active learning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：