P. Hubbard, Ruonan Ou, Tian-Ji Xu, Linqing Luo, H. Nonaka, M. Karrenbach, K. Soga
Distributed acoustic sensing (DAS) is a new technology that is being adopted widely in the geophysics and earth science communities to measure seismic signals propagating over tens of kilometers using an optical fiber. DAS uses the technique of phase‐coherent optical time domain reflectometry (φ‐OTDR) to measure dynamic strain in an optical fiber as small as nε by examining interferences in Rayleigh‐backscattered light. This technology is opening a new field of research of examining very small strains in infrastructure that are much smaller than what is currently able to be measured with the commonly used Brillouin‐based fiber optic sensing technologies. These small strains can be indicative of infrastructure's performance and use level. In this study, a fiber optic strain sensing cable was embedded into an asphalt concrete test road and spatially distributed dynamic road strain was measured during different types of loading. The study's results demonstrate that φ‐OTDR can be used to quantitatively measure strain in roads associated with events as small as a dog walking on the surface. Optical frequency domain reflectometry (OFDR), a widely implemented but less accurate distributed fiber optic strain monitoring technology, was also used along with traditional pavement strain gauges and 3D finite element modeling to validate the φ‐OTDR pavement strain measurements. After validation, φ‐OTDR strain measurements from various events are presented including a vehicle, pedestrian, runner, cyclist, and finally a dog moving along the road. This study serves to demonstrate the deployment of φ‐OTDR to monitor roadway systems.
{"title":"Road deformation monitoring and event detection using asphalt‐embedded distributed acoustic sensing (DAS)","authors":"P. Hubbard, Ruonan Ou, Tian-Ji Xu, Linqing Luo, H. Nonaka, M. Karrenbach, K. Soga","doi":"10.1002/stc.3067","DOIUrl":"https://doi.org/10.1002/stc.3067","url":null,"abstract":"Distributed acoustic sensing (DAS) is a new technology that is being adopted widely in the geophysics and earth science communities to measure seismic signals propagating over tens of kilometers using an optical fiber. DAS uses the technique of phase‐coherent optical time domain reflectometry (φ‐OTDR) to measure dynamic strain in an optical fiber as small as nε by examining interferences in Rayleigh‐backscattered light. This technology is opening a new field of research of examining very small strains in infrastructure that are much smaller than what is currently able to be measured with the commonly used Brillouin‐based fiber optic sensing technologies. These small strains can be indicative of infrastructure's performance and use level. In this study, a fiber optic strain sensing cable was embedded into an asphalt concrete test road and spatially distributed dynamic road strain was measured during different types of loading. The study's results demonstrate that φ‐OTDR can be used to quantitatively measure strain in roads associated with events as small as a dog walking on the surface. Optical frequency domain reflectometry (OFDR), a widely implemented but less accurate distributed fiber optic strain monitoring technology, was also used along with traditional pavement strain gauges and 3D finite element modeling to validate the φ‐OTDR pavement strain measurements. After validation, φ‐OTDR strain measurements from various events are presented including a vehicle, pedestrian, runner, cyclist, and finally a dog moving along the road. This study serves to demonstrate the deployment of φ‐OTDR to monitor roadway systems.","PeriodicalId":22049,"journal":{"name":"Structural Control and Health Monitoring","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2022-08-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82114570","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Shengjun Xu, Ming Hao, Guang-Hui Liu, Yuebo Meng, Jiu-Qiang Han, Ya Shi
Automatic crack detection on concrete surfaces has become increasingly important for the health diagnosis of concrete structures to prevent possible malfunctions or accidents. In this paper, a concrete crack segmentation network based on convolution–deconvolution feature fusion with holistically nested networks is proposed. The proposed network adopts an encoder–decoder structure and uses VGG‐16 as the basic feature extraction network. First, considering the problem that the VGG‐16 network can extract redundant features in the encoding stage, based on the channel attention mechanism, the channel spatial correlation and global information are used to emphasize crack features to remove redundant features. Second, through the convolution–deconvolution feature fusion module, the deep semantic information of the deconvolution is effectively fused with the shallow features of convolution, which effectively improves the semantic crack feature information extracted at each stage of the VGG‐16 network. Finally, based on a multiscale supervised learning mechanism, holistically nested networks are used to fuse the prediction results from different scales, which enhances the network's ability to express linear topological structures and improves the accuracy of crack segmentation. Through a large number of experiments on the Bridge_Crack_Image_Data dataset and CFD dataset, we demonstrate that compared with other deep networks, the proposed network not only achieves better segmentation results for cracks of different widths but is also more robust.
{"title":"Concrete crack segmentation based on convolution–deconvolution feature fusion with holistically nested networks","authors":"Shengjun Xu, Ming Hao, Guang-Hui Liu, Yuebo Meng, Jiu-Qiang Han, Ya Shi","doi":"10.1002/stc.2965","DOIUrl":"https://doi.org/10.1002/stc.2965","url":null,"abstract":"Automatic crack detection on concrete surfaces has become increasingly important for the health diagnosis of concrete structures to prevent possible malfunctions or accidents. In this paper, a concrete crack segmentation network based on convolution–deconvolution feature fusion with holistically nested networks is proposed. The proposed network adopts an encoder–decoder structure and uses VGG‐16 as the basic feature extraction network. First, considering the problem that the VGG‐16 network can extract redundant features in the encoding stage, based on the channel attention mechanism, the channel spatial correlation and global information are used to emphasize crack features to remove redundant features. Second, through the convolution–deconvolution feature fusion module, the deep semantic information of the deconvolution is effectively fused with the shallow features of convolution, which effectively improves the semantic crack feature information extracted at each stage of the VGG‐16 network. Finally, based on a multiscale supervised learning mechanism, holistically nested networks are used to fuse the prediction results from different scales, which enhances the network's ability to express linear topological structures and improves the accuracy of crack segmentation. Through a large number of experiments on the Bridge_Crack_Image_Data dataset and CFD dataset, we demonstrate that compared with other deep networks, the proposed network not only achieves better segmentation results for cracks of different widths but is also more robust.","PeriodicalId":22049,"journal":{"name":"Structural Control and Health Monitoring","volume":"3 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85206033","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Dong‐Hui Yang, Hong Zhou, T. Yi, Hong‐Nan Li, H. Bai
As critical components that transmit internal forces laterally in multigirder bridges, joint degradation can destroy the cooperative working mechanism of a multigirder system and seriously reduce bridge bearing capacity. This research aims to reveal the joint damage effects on lateral internal force transmission and propose a joint damage detection and location method. A spring‐jointed plate model is established to analyze the effects of joint damage, based on which the relationship between the shear forces in the joints and the shear stiffness of the joints can be obtained. Furthermore, a joint damage index is deduced based on the lateral deflection influence lines, and a bridge load testing method is proposed to obtain such influence lines. By using multiple influence lines to jointly solve the damage index, the joint damage in the lateral bridge direction can be accurately detected and located. In addition, by carrying out the damage detection process at several cross sections of the bridge, the joint damage position in the longitudinal bridge direction can also be determined. Finally, a numerical model of an adjacent box girder bridge is illustrated as an example to verify the effectiveness of the damage detection. It can be concluded that the proposed method can effectively identify and locate single or multiple joint damage locations under noise interference, and the joint damage degree can be quantitatively evaluated. This study can provide an effective method to identify the joint damage for adjacent box girder bridges with high accuracy and reliability.
{"title":"Joint deterioration detection based on field‐identified lateral deflection influence lines for adjacent box girder bridges","authors":"Dong‐Hui Yang, Hong Zhou, T. Yi, Hong‐Nan Li, H. Bai","doi":"10.1002/stc.3053","DOIUrl":"https://doi.org/10.1002/stc.3053","url":null,"abstract":"As critical components that transmit internal forces laterally in multigirder bridges, joint degradation can destroy the cooperative working mechanism of a multigirder system and seriously reduce bridge bearing capacity. This research aims to reveal the joint damage effects on lateral internal force transmission and propose a joint damage detection and location method. A spring‐jointed plate model is established to analyze the effects of joint damage, based on which the relationship between the shear forces in the joints and the shear stiffness of the joints can be obtained. Furthermore, a joint damage index is deduced based on the lateral deflection influence lines, and a bridge load testing method is proposed to obtain such influence lines. By using multiple influence lines to jointly solve the damage index, the joint damage in the lateral bridge direction can be accurately detected and located. In addition, by carrying out the damage detection process at several cross sections of the bridge, the joint damage position in the longitudinal bridge direction can also be determined. Finally, a numerical model of an adjacent box girder bridge is illustrated as an example to verify the effectiveness of the damage detection. It can be concluded that the proposed method can effectively identify and locate single or multiple joint damage locations under noise interference, and the joint damage degree can be quantitatively evaluated. This study can provide an effective method to identify the joint damage for adjacent box girder bridges with high accuracy and reliability.","PeriodicalId":22049,"journal":{"name":"Structural Control and Health Monitoring","volume":"61 2","pages":""},"PeriodicalIF":0.0,"publicationDate":"2022-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72560661","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
As one of the most important load‐bearing components of cable‐stayed bridges, the integrity of cables is essential to bridge condition assessment. However, the traditional one‐dimensional damage location method has difficulty determining the spatial location of the damage source on parallel steel wire cables. Acoustic emission (AE) technology is a common means of structural health monitoring, and one of its most beneficial attributes is the ability to localize the damage. For this reason, a three‐dimensional (3D) location algorithm based on the AE signal propagation path is proposed for cable damage localization. First, the propagation path of simulated AE signals in a parallel steel wire bundle (PSWB) was visualized using COMSOL Multiphysics software platform. Then, a 3D localization algorithm suitable for PSWB damage is proposed based on the propagation path and dispersion characteristics of AE waves. After that, experimental verifications were performed, and it was found that the damage locations can be determined accurately using the proposed algorithm. This algorithm helps to localize the damage source with only a few sensors and is crucial for cable protection and replacement.
{"title":"Investigation of wave propagation path and damage source 3D localization in parallel steel wire bundle","authors":"Zhenwen Liu, Shengli Li, Lulu Liu","doi":"10.1002/stc.3051","DOIUrl":"https://doi.org/10.1002/stc.3051","url":null,"abstract":"As one of the most important load‐bearing components of cable‐stayed bridges, the integrity of cables is essential to bridge condition assessment. However, the traditional one‐dimensional damage location method has difficulty determining the spatial location of the damage source on parallel steel wire cables. Acoustic emission (AE) technology is a common means of structural health monitoring, and one of its most beneficial attributes is the ability to localize the damage. For this reason, a three‐dimensional (3D) location algorithm based on the AE signal propagation path is proposed for cable damage localization. First, the propagation path of simulated AE signals in a parallel steel wire bundle (PSWB) was visualized using COMSOL Multiphysics software platform. Then, a 3D localization algorithm suitable for PSWB damage is proposed based on the propagation path and dispersion characteristics of AE waves. After that, experimental verifications were performed, and it was found that the damage locations can be determined accurately using the proposed algorithm. This algorithm helps to localize the damage source with only a few sensors and is crucial for cable protection and replacement.","PeriodicalId":22049,"journal":{"name":"Structural Control and Health Monitoring","volume":"23 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2022-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82177829","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
C. Bernuzzi, C. Rottenbacher, M. Simoncelli, P. Venini
During the Emilia‐Romagna earthquake (2012), a great number of steel racks used to store cheese wheels collapsed, causing a non‐negligible damage to the Italian economy. Therefore, for similar structures that survived and are in service, a deep investigation towards the assessment of their effective safety is required.
{"title":"Modal identification of storage racks for cheese wheels","authors":"C. Bernuzzi, C. Rottenbacher, M. Simoncelli, P. Venini","doi":"10.1002/stc.3052","DOIUrl":"https://doi.org/10.1002/stc.3052","url":null,"abstract":"During the Emilia‐Romagna earthquake (2012), a great number of steel racks used to store cheese wheels collapsed, causing a non‐negligible damage to the Italian economy. Therefore, for similar structures that survived and are in service, a deep investigation towards the assessment of their effective safety is required.","PeriodicalId":22049,"journal":{"name":"Structural Control and Health Monitoring","volume":"47 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2022-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79233181","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Umakanta Meher, Sudhanshu Kumar Mishra, M. R. Sunny
A detection technique to quantify the degree of bolt looseness in metallic bolted structure using electro‐mechanical impedance signatures is proposed. A bolted joint connection of two steel plates and a stiffener is taken as the specimen to be monitored. Loosening of the bolted joints is considered as the damage present in the structure. At first, the electro‐mechanical responses at two piezoelectric transducer locations are measured experimentally for the undamaged and damaged state of the structure. Damage scenarios with single as well as multiple degrees of bolt looseness are considered. Damage features based on root mean square deviation (RMSD) and correlation coefficient (CC) of conductance with respect to the healthy state conductance are extracted. A single hidden layer backpropagation artificial neural network has been trained for detection of bolt looseness from the damage features. Acceptability of the proposed multiple damage detection technique has been observed through few test cases.
{"title":"Impedance‐based looseness detection of bolted joints using artificial neural network: An experimental study","authors":"Umakanta Meher, Sudhanshu Kumar Mishra, M. R. Sunny","doi":"10.1002/stc.3049","DOIUrl":"https://doi.org/10.1002/stc.3049","url":null,"abstract":"A detection technique to quantify the degree of bolt looseness in metallic bolted structure using electro‐mechanical impedance signatures is proposed. A bolted joint connection of two steel plates and a stiffener is taken as the specimen to be monitored. Loosening of the bolted joints is considered as the damage present in the structure. At first, the electro‐mechanical responses at two piezoelectric transducer locations are measured experimentally for the undamaged and damaged state of the structure. Damage scenarios with single as well as multiple degrees of bolt looseness are considered. Damage features based on root mean square deviation (RMSD) and correlation coefficient (CC) of conductance with respect to the healthy state conductance are extracted. A single hidden layer backpropagation artificial neural network has been trained for detection of bolt looseness from the damage features. Acceptability of the proposed multiple damage detection technique has been observed through few test cases.","PeriodicalId":22049,"journal":{"name":"Structural Control and Health Monitoring","volume":"63 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2022-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74464199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Sheng Shi, D. Du, O. Mercan, Erol Kalkan, Shu-guang Wang
Real‐time structural damage detection is one of the main goals of establishing an effective structural health monitoring system. However, due to the lack of training data for possible damage patterns, supervised methods tend to be difficult for such applications. This article therefore proposes a novel unsupervised real‐time damage detection method using machine learning, which consists of a statistical modeling approach using neural networks and a decision‐making process using deep support vector domain description. To choose an optimal window length while extracting damage‐sensitive features, an iterative training strategy is proposed to remove redundant samples from an oversized window. The proposed method is then verified using a simulated dataset from the International Association for Structural Control–American Society of Civil Engineering benchmark and an experimental dataset from shake table tests. The results show that the mean alarm density can be used as an indicator of damage existence and damage levels for the single‐sensor approach. Higher performance of damage detection and lower performance of identifying damage levels are observed for the multi‐sensor approach when the rotational modes are amplified by asymmetric damage patterns. The results of mean false alarm density show that the presented method has a low probability of generating false alarms. The effectiveness of iterative pruning strategy is observed through the visualization of loss function and weights in the neural networks. Finally, the capability of real‐time execution of the proposed damage detection method is investigated and verified. As a result, trained with healthy data only, the proposed method is effective in detecting damage existence and damage levels.
{"title":"A novel unsupervised real‐time damage detection method for structural health monitoring using machine learning","authors":"Sheng Shi, D. Du, O. Mercan, Erol Kalkan, Shu-guang Wang","doi":"10.1002/stc.3042","DOIUrl":"https://doi.org/10.1002/stc.3042","url":null,"abstract":"Real‐time structural damage detection is one of the main goals of establishing an effective structural health monitoring system. However, due to the lack of training data for possible damage patterns, supervised methods tend to be difficult for such applications. This article therefore proposes a novel unsupervised real‐time damage detection method using machine learning, which consists of a statistical modeling approach using neural networks and a decision‐making process using deep support vector domain description. To choose an optimal window length while extracting damage‐sensitive features, an iterative training strategy is proposed to remove redundant samples from an oversized window. The proposed method is then verified using a simulated dataset from the International Association for Structural Control–American Society of Civil Engineering benchmark and an experimental dataset from shake table tests. The results show that the mean alarm density can be used as an indicator of damage existence and damage levels for the single‐sensor approach. Higher performance of damage detection and lower performance of identifying damage levels are observed for the multi‐sensor approach when the rotational modes are amplified by asymmetric damage patterns. The results of mean false alarm density show that the presented method has a low probability of generating false alarms. The effectiveness of iterative pruning strategy is observed through the visualization of loss function and weights in the neural networks. Finally, the capability of real‐time execution of the proposed damage detection method is investigated and verified. As a result, trained with healthy data only, the proposed method is effective in detecting damage existence and damage levels.","PeriodicalId":22049,"journal":{"name":"Structural Control and Health Monitoring","volume":"19 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2022-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91554577","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
G. Fu, Hao Wang, Yanqing Zhao, Zhanqiang Yu, Qiang Li
In order to select the optimal treatment strategy for cracked pavements, the cracking conditions should be accurately investigated and evaluated. In this study, the effects of longitudinal cracking on falling weight deflectometer (FWD) deflections were investigated, and a rapid and non‐destructive approach was accordingly proposed to evaluate the longitudinal cracking severity using FWD data for semi‐rigid pavements. 3D finite element models were developed to simulate various intact and cracked pavements to compute the surface deflections under FWD loading. Two cracking types, namely, cracking in asphalt concrete layer (AC cracking) and cracking in both AC and cement‐treated base layers (AC + CTB cracking), were considered. In most cases analyzed, the deflections of cracked pavements are greater than those of intact pavements, and they are only slightly smaller than those of intact pavements in other cases. The effects of longitudinal cracking on deflections increase with increasing crack width and decreasing distance between the crack and the loading center, and longitudinal cracking generally has greater influences on the pavement with a thicker AC layer and weaker subgrade. The effects of AC + CTB cracking on deflections are significantly greater than AC cracking, especially for the cracks near the loading center, and the influences of both AC cracking and AC + CTB cracking are negligible when the deflections are measured more than 1.8 m away from the crack. Accordingly, a rapid and non‐destructive approach was proposed to distinguish the AC cracking and AC + CTB cracking using FWD data for semi‐rigid pavements.
{"title":"Non‐destructive evaluation of longitudinal cracking in semi‐rigid asphalt pavements using FWD deflection data","authors":"G. Fu, Hao Wang, Yanqing Zhao, Zhanqiang Yu, Qiang Li","doi":"10.1002/stc.3050","DOIUrl":"https://doi.org/10.1002/stc.3050","url":null,"abstract":"In order to select the optimal treatment strategy for cracked pavements, the cracking conditions should be accurately investigated and evaluated. In this study, the effects of longitudinal cracking on falling weight deflectometer (FWD) deflections were investigated, and a rapid and non‐destructive approach was accordingly proposed to evaluate the longitudinal cracking severity using FWD data for semi‐rigid pavements. 3D finite element models were developed to simulate various intact and cracked pavements to compute the surface deflections under FWD loading. Two cracking types, namely, cracking in asphalt concrete layer (AC cracking) and cracking in both AC and cement‐treated base layers (AC + CTB cracking), were considered. In most cases analyzed, the deflections of cracked pavements are greater than those of intact pavements, and they are only slightly smaller than those of intact pavements in other cases. The effects of longitudinal cracking on deflections increase with increasing crack width and decreasing distance between the crack and the loading center, and longitudinal cracking generally has greater influences on the pavement with a thicker AC layer and weaker subgrade. The effects of AC + CTB cracking on deflections are significantly greater than AC cracking, especially for the cracks near the loading center, and the influences of both AC cracking and AC + CTB cracking are negligible when the deflections are measured more than 1.8 m away from the crack. Accordingly, a rapid and non‐destructive approach was proposed to distinguish the AC cracking and AC + CTB cracking using FWD data for semi‐rigid pavements.","PeriodicalId":22049,"journal":{"name":"Structural Control and Health Monitoring","volume":"51 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2022-07-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85905515","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yuxuan Zhang, Xiao-yang Wang, Z. Ding, Yao Du, Y. Xia
Structural health monitoring (SHM) systems generate a massive amount of sensing data. On one hand, sensor faults may cause the measurement data to have low fidelity. On the other hand, extreme events, such as typhoons or earthquakes, may cause the monitoring data look “abnormal.” These abnormal data, however, are closely related to the structural safety condition and require special attention. This study proposes an automatic and efficient anomaly detection methodology based on support vector data description (SVDD) to simultaneously detect anomalies caused by sensor faults and extreme events. The SVDD trained by a single pattern can divide the feature space into one‐versus‐the rest. Several decision boundaries are defined to enclose normal data and common sensor fault patterns, forming an equivalent multi‐class classifier to classify common sensor fault types and detect unknown patterns. Next, multiple sensor faults and extreme events are separated from the unknown patterns. Multi‐label data are detected based on the local features, while extreme events are recognized by the correlation of different sensors. The proposed method is finally applied to datasets collected from two SHM systems. Results show that the sensor anomalies in the systems are detected with high efficiency and accuracy, and extreme events are separated as a special pattern from the normal, common abnormal, and unknown patterns.
{"title":"Anomaly detection of sensor faults and extreme events based on support vector data description","authors":"Yuxuan Zhang, Xiao-yang Wang, Z. Ding, Yao Du, Y. Xia","doi":"10.1002/stc.3047","DOIUrl":"https://doi.org/10.1002/stc.3047","url":null,"abstract":"Structural health monitoring (SHM) systems generate a massive amount of sensing data. On one hand, sensor faults may cause the measurement data to have low fidelity. On the other hand, extreme events, such as typhoons or earthquakes, may cause the monitoring data look “abnormal.” These abnormal data, however, are closely related to the structural safety condition and require special attention. This study proposes an automatic and efficient anomaly detection methodology based on support vector data description (SVDD) to simultaneously detect anomalies caused by sensor faults and extreme events. The SVDD trained by a single pattern can divide the feature space into one‐versus‐the rest. Several decision boundaries are defined to enclose normal data and common sensor fault patterns, forming an equivalent multi‐class classifier to classify common sensor fault types and detect unknown patterns. Next, multiple sensor faults and extreme events are separated from the unknown patterns. Multi‐label data are detected based on the local features, while extreme events are recognized by the correlation of different sensors. The proposed method is finally applied to datasets collected from two SHM systems. Results show that the sensor anomalies in the systems are detected with high efficiency and accuracy, and extreme events are separated as a special pattern from the normal, common abnormal, and unknown patterns.","PeriodicalId":22049,"journal":{"name":"Structural Control and Health Monitoring","volume":"23 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2022-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81762269","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Monitoring the deformation or displacement response of buildings is critical for structural safety. Recently, the development of computer vision has led to extensive research on the application of vision‐based measurements in the structural monitoring. This enables the use of urban surveillance video cameras, which are widely installed and can produce numerous images and videos of urban scenes to measure the structural displacement. However, the structural displacement measurement may be inaccurate owing to the limited hardware resolution of the surveillance video cameras or the long distance from the cameras to the monitored targets. To this end, this study proposes a method to improve the displacement measurement accuracy using a deep learning super‐resolution model based on generative adversarial networks. The proposed method achieves texture detail enhancement of low‐resolution images or videos by supplementing high‐resolution photographs of the target, thus improving the accuracy of the vision‐based displacement measurement. The proposed method shows good accuracy and stability in both the static and dynamic experimental validations compared with the original low‐resolution images/video and interpolation‐based super‐resolution images/video. In conclusion, the proposed method can support the displacement measurement of buildings and infrastructures based on urban surveillance video cameras.
{"title":"Vision‐based displacement measurement enhanced by super‐resolution using generative adversarial networks","authors":"Chujin Sun, Donglian Gu, Yi Zhang, Xinzheng Lu","doi":"10.1002/stc.3048","DOIUrl":"https://doi.org/10.1002/stc.3048","url":null,"abstract":"Monitoring the deformation or displacement response of buildings is critical for structural safety. Recently, the development of computer vision has led to extensive research on the application of vision‐based measurements in the structural monitoring. This enables the use of urban surveillance video cameras, which are widely installed and can produce numerous images and videos of urban scenes to measure the structural displacement. However, the structural displacement measurement may be inaccurate owing to the limited hardware resolution of the surveillance video cameras or the long distance from the cameras to the monitored targets. To this end, this study proposes a method to improve the displacement measurement accuracy using a deep learning super‐resolution model based on generative adversarial networks. The proposed method achieves texture detail enhancement of low‐resolution images or videos by supplementing high‐resolution photographs of the target, thus improving the accuracy of the vision‐based displacement measurement. The proposed method shows good accuracy and stability in both the static and dynamic experimental validations compared with the original low‐resolution images/video and interpolation‐based super‐resolution images/video. In conclusion, the proposed method can support the displacement measurement of buildings and infrastructures based on urban surveillance video cameras.","PeriodicalId":22049,"journal":{"name":"Structural Control and Health Monitoring","volume":"57 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2022-07-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89480086","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：