Research

Jump to: Review

14 pages, 5025 KiB

Open AccessArticle

Interpretability Analysis of Convolutional Neural Networks for Crack Detection

by Jie Wu, Yongjin He, Chengyu Xu, Xiaoping Jia, Yule Huang, Qianru Chen, Chuyue Huang, Armin Dadras Eslamlou and Shiping Huang

Buildings 2023, 13(12), 3095; https://doi.org/10.3390/buildings13123095 - 13 Dec 2023

Cited by 2 | Viewed by 832

Abstract

Crack detection is an important task in bridge health monitoring, and related detection methods have gradually shifted from traditional manual methods to intelligent approaches with convolutional neural networks (CNNs) in recent years. Due to the opaque process of training and operating CNNs, if [...] Read more.

Crack detection is an important task in bridge health monitoring, and related detection methods have gradually shifted from traditional manual methods to intelligent approaches with convolutional neural networks (CNNs) in recent years. Due to the opaque process of training and operating CNNs, if the learned features for identifying cracks in the network are not evaluated, it may lead to safety risks. In this study, to evaluate the recognition basis of different crack detection networks; several crack detection CNNs are trained using the same training conditions. Afterwards, several crack images are used to construct a dataset, which are used to interpret and analyze the trained networks and obtain the learned features for identifying cracks. Additionally, a crack identification performance criterion based on interpretability analysis is proposed. Finally, a training framework is introduced based on the issues reflected in the interpretability analysis. Full article

(This article belongs to the Special Issue Soft Computing for Structural Health Monitoring)

► Show Figures

Figure 1

26 pages, 4143 KiB

Open AccessArticle

Multimodal Deep Neural Network-Based Sensor Data Anomaly Diagnosis Method for Structural Health Monitoring

by Xingzhong Nong, Xu Luo, Shan Lin, Yanmei Ruan and Xijun Ye

Buildings 2023, 13(8), 1976; https://doi.org/10.3390/buildings13081976 - 02 Aug 2023

Cited by 2 | Viewed by 972

Abstract

Due to sensor failure, noise interference and other factors, the data collected in the structural health monitoring (SHM) system will show a variety of abnormal patterns, which will bring great uncertainty to the structural safety assessment. This paper proposes an automatic data anomaly [...] Read more.

Due to sensor failure, noise interference and other factors, the data collected in the structural health monitoring (SHM) system will show a variety of abnormal patterns, which will bring great uncertainty to the structural safety assessment. This paper proposes an automatic data anomaly diagnosis method for SHM based on a multimodal deep neural network. In order to improve the detection accuracy, both two-dimensional and one-dimensional features of the sensor data are fused in the multimodal deep neural network. The network consists of two convolutional neural network (CNN) channels, one a 2D-CNN channel for extracting time–frequency features of sensor data and the other a 1D-CNN channel for extracting raw one-dimensional features of sensor data. After convolution and pooling operations for the sensor data by the 2D channel and 1D channel separately, the two types of extracted features are flattened into one-dimensional vectors and concatenated at the concatenation layer. The concatenated vector is then fed into fully connected layers for final SHM data anomaly classification. In order to evaluate the reliability of the proposed method, the monitored data lasting for one month of a long-span cable-stayed bridge were used for training, validation, and testing. Six types of training conditions (missing, minor, outlier, over-range oscillation, trend, and drift) are studied and analyzed to address the issue of imbalanced training data. With an accuracy rate of 95.10%, the optimal model demonstrates the effectiveness and capability of the proposed method. The proposed method shows a promising future as a reliable AI-assisted digital tool for safety assessment in structural health monitoring systems. Full article

(This article belongs to the Special Issue Soft Computing for Structural Health Monitoring)

► Show Figures

Figure 1

19 pages, 5876 KiB

Open AccessArticle

Research on One-to-Two Internal Resonance of Sling and Beam of Suspension Sling–Beam System

by Lixiong Gu, Chunguang Dong, Yi Zhang, Xiaoxia Zhen, Guiyuan Liu and Jianyi Ji

Buildings 2023, 13(5), 1319; https://doi.org/10.3390/buildings13051319 - 18 May 2023

Viewed by 718

Abstract

An approach is presented to investigate the 1:2 internal resonance of the sling and beam of a suspension sling–beam system. The beam was taken as the geometrically linear Euler beam, and the sling was considered to be geometrically nonlinear. The dynamic equilibrium equation [...] Read more.

An approach is presented to investigate the 1:2 internal resonance of the sling and beam of a suspension sling–beam system. The beam was taken as the geometrically linear Euler beam, and the sling was considered to be geometrically nonlinear. The dynamic equilibrium equation of the structures was derived using the modal superposition method, the D’Alembert principle and the Hamilton principle. The nonlinear dynamic equilibrium equations of free vibration and forced oscillation were solved by the multiple-scales method. We derived the first approximation solutions for the single-modal motion of the system. Numerical examples are provided to verify the correctness of formula derivation and obtain the amplitude–time response of free vibration, the primary resonance response, the amplitude–frequency response, and the amplitude–force response of forced oscillation. According to the analysis, it is evident that the combination system exhibits robust nonlinear coupling properties due to the presence of internal resonance, which are useful for engineering design. Full article

(This article belongs to the Special Issue Soft Computing for Structural Health Monitoring)

► Show Figures

Figure 1

23 pages, 8076 KiB

Open AccessArticle

Numerical Analysis on the Impact Effect of Cable Breaking for a New Type Arch Bridge

by Jianhong Huo, Yonghui Huang, Jialin Wang and Qiye Zhuo

Buildings 2023, 13(3), 753; https://doi.org/10.3390/buildings13030753 - 13 Mar 2023

Viewed by 1314

Abstract

Taking Haixin Bridge as an example, the structural response of a new type arch bridge composed of an inclined arch and a curved beam under cable breaking is analyzed numerically. The cable breaking at different positions, different numbers of broken cables and different [...] Read more.

Taking Haixin Bridge as an example, the structural response of a new type arch bridge composed of an inclined arch and a curved beam under cable breaking is analyzed numerically. The cable breaking at different positions, different numbers of broken cables and different ways of breaking are modeled and calculated, and the remaining cables’ internal force and main girder’s deflection are selected as research indexes to evaluate the degree of impact effect of broken cables on the bridge. The numerical results show that (1) duration time of cable breaking is an important factor affecting the impact effect of the bridge, when the cable breaking time is less than 1% of the first order natural vibration period of the structure, the dynamic response caused by cable breaking no longer variates with time; (2) for the cables adjacent to the breaking cable at equal distances, the cable with a shorter length will carry more released force of breaking cable than the longer, and the impact effect is more significant; (3) the dynamic response of displacement and cable force caused by cables at different locations are different, a cable located in the L/4 arch rib area suddenly breaking shows the largest dynamic response; (4) it is feasible to take the dynamic amplification factor (DAF) of cable force and the main girder’s deflection as 2, but it is unsafe to take the DAF of the arch rib’s deflection as 2; (5) the dynamic response of multiple cables breaking at the same time cannot be simplified as a linear superposition of single cable breaks one by one, and the amplification effect becomes more significant with the increase of the number of broken cables. These conclusions can provide guidance for structural safety assessment of similar arch bridges after cable breakage. Full article

(This article belongs to the Special Issue Soft Computing for Structural Health Monitoring)

► Show Figures

Figure 1

14 pages, 5815 KiB

Open AccessArticle

An FPGA-Based Laser Virtual Scale Method for Structural Crack Measurement

by Miaomiao Yuan, Zhuneng Fang, Peng Xiao, Ruijin Tong, Min Zhang and Yule Huang

Buildings 2023, 13(1), 261; https://doi.org/10.3390/buildings13010261 - 16 Jan 2023

Cited by 1 | Viewed by 1683

Abstract

Real-time systems for measuring structural cracks are of great significance due to their computational and cost efficacy, inherent hazards, and detection discrepancies associated with the manual visual assessment of structures. The precision and effectiveness of image measurement approaches increased their applications in vast [...] Read more.

Real-time systems for measuring structural cracks are of great significance due to their computational and cost efficacy, inherent hazards, and detection discrepancies associated with the manual visual assessment of structures. The precision and effectiveness of image measurement approaches increased their applications in vast regions. This article proposes a field-programmable gate array (FPGA)-based laser virtual scale algorithm for noncontact real-time measurement of structural crack images. The device first sends two parallel beams and then applies image processing techniques, including de-noising with median and morphological filtering, as well as Sobel-operator-based edge extraction, to process and localize the light spots. Afterwards, it acquires the scale of the pixel distance to the physical distance and then derives the actual size of the crack. By processing and positioning, the FPGA acquires the scale of the pixel distance to the physical space and then derives the actual size of the crack. The experimental study on crack measurements demonstrates that the proposed technique has precise and reliable results. The error rate is approximately 2.47%, sufficient to meet measurement accuracy criteria. Moreover, experimental results suggest that the processing time for one frame using an FPGA is about 54 ms, and that the hardware acceleration provided using an FPGA is approximately 120 times that of a PC, allowing for real-time operation. The proposed method is a simple and computationally efficient tool with better efficacy for noncontact measurements. Full article

(This article belongs to the Special Issue Soft Computing for Structural Health Monitoring)

► Show Figures

Figure 1

13 pages, 3673 KiB

Open AccessArticle

Research on Parametric Vibration of a Steel Truss Corridor under Pedestrians Excitation Considering the Time-Delay Effect

by Zhou Chen, Jiahao Wen, Deyuan Deng, Wei Dai, Siyuan Chen, Hanwen Lu and Lingfei Liu

Buildings 2023, 13(1), 98; https://doi.org/10.3390/buildings13010098 - 30 Dec 2022

Cited by 1 | Viewed by 1131

Abstract

With the development of the Steel Truss Corridor (STC) toward long-span and gentle development, human-induced vibration often causes large lateral vibration problems and time-delay effects of the STC, which will have a non-negligible impact on the dynamic performance of the STC. In this [...] Read more.

With the development of the Steel Truss Corridor (STC) toward long-span and gentle development, human-induced vibration often causes large lateral vibration problems and time-delay effects of the STC, which will have a non-negligible impact on the dynamic performance of the STC. In this paper, the parametric vibration model proposed by Piccardo is improved, and the nonlinear dynamic equation of the STC is established considering the longitudinal–lateral walking force coupled parametric vibration with the time-delay effect. Taking the Millennium STC as an example, the mechanism of lateral vibration under the time-delay effect is discussed by the numerical calculation method, and the influence of the time-delay effect on its dynamic response is analyzed. The results show that: considering the time-delay effect, when the frequency ratio θ/2Ω = 1, the value of the time-delay coefficient has no effect on the critical number of STCs in the parametric resonance region. As θ/2Ω moves away from 1, the more significant the effect. When the STC begins to excite the parametric resonance phenomenon, the existence of the time-delay effect will change the time for the STC to reach a stable amplitude and suppress the lateral vibration of the STC. When the STC generates parametric vibration, the value of the time-delay coefficient has no effect on the nonlinear dynamic response of the STC. For STCs in both the nonparametric resonance region and the critical region, there is a pair of staggered critical bifurcation time-delay coefficients, which increase or decrease the vibration response. Full article

(This article belongs to the Special Issue Soft Computing for Structural Health Monitoring)

► Show Figures

Figure 1

17 pages, 4716 KiB

Open AccessArticle

Evaluation of the Cable Force by Frequency Method for the Hybrid Boundary between the Ear Plate and the Anchor Plate

by Yufeng Xu, Yunfei Xie, Si Chen and Mengyang Zhu

Buildings 2022, 12(11), 1853; https://doi.org/10.3390/buildings12111853 - 02 Nov 2022

Cited by 1 | Viewed by 1098

Abstract

For the quick and accurate determination of the cable force under the hybrid boundary between the ear plate and the anchor plate, this paper proposes a method for identifying the boundary conditions and bending stiffness of the cables based on the frequency method. [...] Read more.

For the quick and accurate determination of the cable force under the hybrid boundary between the ear plate and the anchor plate, this paper proposes a method for identifying the boundary conditions and bending stiffness of the cables based on the frequency method. First, the influence of the parameters, namely, cable boundary conditions, bending stiffness, inclination angle, and linear density, on the cable force was analyzed. Next, the actual bending stiffness of the cable under the boundary conditions of the upper hinged support and the lower solid support and the solid support at both ends were determined, and the cable force in the two states was calculated. Finally, the hydraulic jack tension and the cable force in the two states were compared and analyzed to define the actual boundary conditions and bending stiffness of the cable. The results show that the boundary conditions, bending stiffness, and linear density of the cable considerably influence the cable force, whereas the inclination angle has a negligible effect on the cable force. The bending stiffness under the two boundary conditions of the cable were both 0.12Eimax, and the calculated value of the cable force under the solid support boundary condition at both ends was consistent with the value of the hydraulic jack tension. Full article

(This article belongs to the Special Issue Soft Computing for Structural Health Monitoring)

► Show Figures

Graphical abstract

Review

Jump to: Research

28 pages, 3532 KiB

Open AccessReview

Artificial-Neural-Network-Based Surrogate Models for Structural Health Monitoring of Civil Structures: A Literature Review

by Armin Dadras Eslamlou and Shiping Huang

Buildings 2022, 12(12), 2067; https://doi.org/10.3390/buildings12122067 - 25 Nov 2022

Cited by 10 | Viewed by 2776

Abstract

It is often computationally expensive to monitor structural health using computer models. This time-consuming process can be relieved using surrogate models, which provide cheap-to-evaluate metamodels to replace the original expensive models. Because of their high accuracy, simplicity, and efficiency, Artificial Neural Networks (ANNs) [...] Read more.

It is often computationally expensive to monitor structural health using computer models. This time-consuming process can be relieved using surrogate models, which provide cheap-to-evaluate metamodels to replace the original expensive models. Because of their high accuracy, simplicity, and efficiency, Artificial Neural Networks (ANNs) have gained considerable attention in this area. This paper reviews the application of ANNs as surrogates for structural health monitoring in the literature. Moreover, the review contains fundamental information, detailed discussions, wide comparisons, and suggestions for future research. Surrogates in this literature review are divided into parametric and nonparametric models. In the past, nonparametric models dominated this field, but parametric models have gained popularity in the recent decade. A parametric surrogate is commonly supplied with metaheuristic algorithms, and can provide high levels of identification. Recurrent networks, instead of traditional ANNs, have also become increasingly popular for nonparametric surrogates. Full article

(This article belongs to the Special Issue Soft Computing for Structural Health Monitoring)

► Show Figures

Figure 1

Journal Menu

Journal Browser

Soft Computing for Structural Health Monitoring

Share This Special Issue

Special Issue Editors

Special Issue Information

Keywords

Published Papers (8 papers)

Research

Review

Further Information

Guidelines

MDPI Initiatives

Follow MDPI