C. Lindley, T. Rogers, R. Dwyer-Joyce, N. Dervilis, K. Worden
In structural health monitoring (SHM) and condition monitoring (CM) applications, the expense of testing programmes may be too high to obtain adequate datasets. When limited by the number of available data samples, one may rely on dimensional reduction methods to proceed with a meaningful statistical and probabilistic analysis. In this work, some state-of-the-art dimensionality-reduction techniques were investigated as part of a simple ball-bearing damage detection problem. A variational auto-encoder (VAE) was compared to other methods, based on their capability to generate low-dimensional representations of the data. Unlike other common alternatives, such as principal component analysis (PCA) or auto-encoding (AE) networks, the VAE introduces a probabilistic framework via the latent embeddings. A well-defined distribution is thereby constructed on the latent variables, making the transformed dataset an optimal one for subsequent pattern recognition analysis. The results demonstrated an increase in classification performance given the low-dimensional representation generated by the VAE.
{"title":"ON THE APPLICATION OF VARIATIONAL AUTO ENCODERS (VAE) FOR DAMAGE DETECTION IN ROLLING ELEMENT BEARINGS","authors":"C. Lindley, T. Rogers, R. Dwyer-Joyce, N. Dervilis, K. Worden","doi":"10.12783/shm2021/36281","DOIUrl":"https://doi.org/10.12783/shm2021/36281","url":null,"abstract":"In structural health monitoring (SHM) and condition monitoring (CM) applications, the expense of testing programmes may be too high to obtain adequate datasets. When limited by the number of available data samples, one may rely on dimensional reduction methods to proceed with a meaningful statistical and probabilistic analysis. In this work, some state-of-the-art dimensionality-reduction techniques were investigated as part of a simple ball-bearing damage detection problem. A variational auto-encoder (VAE) was compared to other methods, based on their capability to generate low-dimensional representations of the data. Unlike other common alternatives, such as principal component analysis (PCA) or auto-encoding (AE) networks, the VAE introduces a probabilistic framework via the latent embeddings. A well-defined distribution is thereby constructed on the latent variables, making the transformed dataset an optimal one for subsequent pattern recognition analysis. The results demonstrated an increase in classification performance given the low-dimensional representation generated by the VAE.","PeriodicalId":180083,"journal":{"name":"Proceedings of the 13th International Workshop on Structural Health Monitoring","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114255689","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}
Mahyar Kazemian, Sajad Nikdel, Mehrnaz Mohammadesmaeili, V. Nik, K. Zandi
Environmental loads, such as wind and river flow, play an essential role in the structural design and structural assessment of long-span bridges. Climate change and extreme climatic events are threats to the reliability and safety of the transport network. This has led to a growing demand for digital twin models to investigate the resilience of bridges under extreme climate conditions. Kalix bridge, constructed over the Kalix river in Sweden in 1956, is used as a testbed in this context. The bridge structure, made of posttensioned concrete, consists of five spans, with the longest one being 94 m. In this study, aerodynamic characteristics and extreme values of numerical wind simulation such as surface pressure are obtained by using Spalart-Allmaras Delayed Detached Eddy Simulation (DDES) as a hybrid RANS-LES turbulence approach which is both practical and computationally efficient for near-wall mesh density imposed by the LES method. Surface wind pressure is obtained for three extreme climate scenarios, including extreme windy weather, extremely cold weather, and design value for a 3000-year return period. The result indicates significant differences in surface wind pressure due to time layers coming from transient wind flow simulation. In order to assess the structural performance under the critical wind scenario, the highest value of surface pressure for each scenario is considered. Also, a hydrodynamic study is conducted on the bridge pillars, in which the river flow is simulated using the VOF method, and the water movement process around the pillars is examined transiently and at different times. The surface pressure applied by the river flow with the highest recorded volumetric flow is calculated on each of the pier surfaces. In simulating the river flow, information and weather conditions recorded in the past periods have been used. The results show that the surface pressure at the time when the river flow hit the pillars is much higher than in subsequent times. This amount of pressure can be used as a critical load in fluid-structure interaction (FSI) calculations. Finally, for both sections, the wind surface pressure, the velocity field with respect to auxiliary probe lines, the water circumferential motion contours around the pillars, and the pressure diagram on them are reported in different timesteps.
{"title":"KALIX BRIDGE DIGITAL TWIN—STRUCTURAL LOADS FROM FUTURE EXTREME CLIMATE EVENTS","authors":"Mahyar Kazemian, Sajad Nikdel, Mehrnaz Mohammadesmaeili, V. Nik, K. Zandi","doi":"10.12783/shm2021/36323","DOIUrl":"https://doi.org/10.12783/shm2021/36323","url":null,"abstract":"Environmental loads, such as wind and river flow, play an essential role in the structural design and structural assessment of long-span bridges. Climate change and extreme climatic events are threats to the reliability and safety of the transport network. This has led to a growing demand for digital twin models to investigate the resilience of bridges under extreme climate conditions. Kalix bridge, constructed over the Kalix river in Sweden in 1956, is used as a testbed in this context. The bridge structure, made of posttensioned concrete, consists of five spans, with the longest one being 94 m. In this study, aerodynamic characteristics and extreme values of numerical wind simulation such as surface pressure are obtained by using Spalart-Allmaras Delayed Detached Eddy Simulation (DDES) as a hybrid RANS-LES turbulence approach which is both practical and computationally efficient for near-wall mesh density imposed by the LES method. Surface wind pressure is obtained for three extreme climate scenarios, including extreme windy weather, extremely cold weather, and design value for a 3000-year return period. The result indicates significant differences in surface wind pressure due to time layers coming from transient wind flow simulation. In order to assess the structural performance under the critical wind scenario, the highest value of surface pressure for each scenario is considered. Also, a hydrodynamic study is conducted on the bridge pillars, in which the river flow is simulated using the VOF method, and the water movement process around the pillars is examined transiently and at different times. The surface pressure applied by the river flow with the highest recorded volumetric flow is calculated on each of the pier surfaces. In simulating the river flow, information and weather conditions recorded in the past periods have been used. The results show that the surface pressure at the time when the river flow hit the pillars is much higher than in subsequent times. This amount of pressure can be used as a critical load in fluid-structure interaction (FSI) calculations. Finally, for both sections, the wind surface pressure, the velocity field with respect to auxiliary probe lines, the water circumferential motion contours around the pillars, and the pressure diagram on them are reported in different timesteps.","PeriodicalId":180083,"journal":{"name":"Proceedings of the 13th International Workshop on Structural Health Monitoring","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122908133","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}
Antoni Lis, Micah Sweeney, M. Samotyj, Artur ARTUR HANC
Machinery monitoring is typically applied to a single machine based on sensor integration and data analysis. Such an approach to a set of machines operating in similar conditions allows for a multivariate analysis for condition monitoring based on a single machine as well as based on group analysis. This paper describes an Industrial Internet-of-Thing (IIoT) concept for condition monitoring of machinery population based on water pumps. The first part provides an introduction to unsupervised anomaly detection based on population modeling with using features calculated from the: mechanical (based on vibration sensors), electrical (voltage and current signals collected from electric motors that drive monitored pumps) and operational processes (such as pressures, flows) signals. Finally, the preliminary results from laboratory testing and demonstration at a wastewater processing plant are presented.
{"title":"POPULATION BASED PUMPS MONITORING AND BENCHMARKING USING IOT AND EDGE ML LEARNING METHODS","authors":"Antoni Lis, Micah Sweeney, M. Samotyj, Artur ARTUR HANC","doi":"10.12783/shm2021/36283","DOIUrl":"https://doi.org/10.12783/shm2021/36283","url":null,"abstract":"Machinery monitoring is typically applied to a single machine based on sensor integration and data analysis. Such an approach to a set of machines operating in similar conditions allows for a multivariate analysis for condition monitoring based on a single machine as well as based on group analysis. This paper describes an Industrial Internet-of-Thing (IIoT) concept for condition monitoring of machinery population based on water pumps. The first part provides an introduction to unsupervised anomaly detection based on population modeling with using features calculated from the: mechanical (based on vibration sensors), electrical (voltage and current signals collected from electric motors that drive monitored pumps) and operational processes (such as pressures, flows) signals. Finally, the preliminary results from laboratory testing and demonstration at a wastewater processing plant are presented.","PeriodicalId":180083,"journal":{"name":"Proceedings of the 13th International Workshop on Structural Health Monitoring","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126995366","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}
Yujin Park, Yingjun Zhao Dubuc, Amy Slider, P. Sessoms, J. Fraser, K. Loh
Lateral ankle sprains cost billions of dollars in medical expenses annually and frequently result in long-term functional decline and a diminished health-related quality of life. While ankle braces have been shown to be effective in prophylaxis of subsequent ankle sprains, current braces are either too stiff and affect normal gait or too flexible and provide insufficient support during high-intensity activities. In this study, we proposed an adaptive ankle brace design that employs dynamically variable stiffness components to provide minimum support under normal gait movements and maximum rigidity under large ranges of motion. To achieve these unique properties, a honeycomb geometry was designed and three dimensionally printed with thermoplastic polyurethane to exhibit nonlinear, strain-stiffening, elastic behavior. We conducted a series of tensile load tests on different honeycomb unit cell configurations. First, the influence of unit cell designs on their mechanical strength and force-strain profiles was characterized. Second, experimentally calibrated finite element models of individual components simulated the mechanical response of the geometry, which were then used to optimize the geometrical parameters of the honeycomb shape (i.e., ring size, length of lateral elements, and thickness). The results identified promising design parameters for these honeycomb geometries that could be used to realize next-generation adaptive ankle braces.
{"title":"VARIABLE STIFFNESS HONEYCOMB METAMATERIALS FOR ADAPTIVE ANKLE BRACE DESIGN","authors":"Yujin Park, Yingjun Zhao Dubuc, Amy Slider, P. Sessoms, J. Fraser, K. Loh","doi":"10.12783/shm2021/36268","DOIUrl":"https://doi.org/10.12783/shm2021/36268","url":null,"abstract":"Lateral ankle sprains cost billions of dollars in medical expenses annually and frequently result in long-term functional decline and a diminished health-related quality of life. While ankle braces have been shown to be effective in prophylaxis of subsequent ankle sprains, current braces are either too stiff and affect normal gait or too flexible and provide insufficient support during high-intensity activities. In this study, we proposed an adaptive ankle brace design that employs dynamically variable stiffness components to provide minimum support under normal gait movements and maximum rigidity under large ranges of motion. To achieve these unique properties, a honeycomb geometry was designed and three dimensionally printed with thermoplastic polyurethane to exhibit nonlinear, strain-stiffening, elastic behavior. We conducted a series of tensile load tests on different honeycomb unit cell configurations. First, the influence of unit cell designs on their mechanical strength and force-strain profiles was characterized. Second, experimentally calibrated finite element models of individual components simulated the mechanical response of the geometry, which were then used to optimize the geometrical parameters of the honeycomb shape (i.e., ring size, length of lateral elements, and thickness). The results identified promising design parameters for these honeycomb geometries that could be used to realize next-generation adaptive ankle braces.","PeriodicalId":180083,"journal":{"name":"Proceedings of the 13th International Workshop on Structural Health Monitoring","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127641550","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}
The initiation and development of microcracks introduced by heating or fire plays a critical role in the stability and durability of concrete structure. However, the concealment of microcracks and the nonlinearity of concrete materials make it difficult to appropriately evaluate the size and extension state of microcracks inside the thermal damaged concrete. In this paper, broadband frequency excitation instead of traditional dual-frequency excitation is utilized to excite the thermal damaged concrete, and the generated ultrasonic modulated signal reflects the micro damage state. The concept of damage index (DI) based on the sideband peak count (SPC) is proposed to quantitatively describe the variation characteristics of modulated signals. The results show that the peak value of DI based on broadband frequency coupling of nonlinear ultrasonic modulation method reflects the generation and development of microcracks in thermal damaged concrete. The peak value of DI increases sensitively with the increasing of water-cement ratio, fine-coarse aggregate ratio, and the heating temperature. Meanwhile, the statistical relationships of the peak value of DI with the residual strength and the area ratio of microcracks in thermal damaged concrete are established respectively.
{"title":"MICROCRACKS DETECTION OF THERMAL DAMAGED CONCRETE WITH NONLINEAR ULTRASONIC MODULATION BASED ON BROAD BAND FREQUENCY COUPLING","authors":"Ying Xu, Heyong Zhang, Huan Liu","doi":"10.12783/shm2021/36360","DOIUrl":"https://doi.org/10.12783/shm2021/36360","url":null,"abstract":"The initiation and development of microcracks introduced by heating or fire plays a critical role in the stability and durability of concrete structure. However, the concealment of microcracks and the nonlinearity of concrete materials make it difficult to appropriately evaluate the size and extension state of microcracks inside the thermal damaged concrete. In this paper, broadband frequency excitation instead of traditional dual-frequency excitation is utilized to excite the thermal damaged concrete, and the generated ultrasonic modulated signal reflects the micro damage state. The concept of damage index (DI) based on the sideband peak count (SPC) is proposed to quantitatively describe the variation characteristics of modulated signals. The results show that the peak value of DI based on broadband frequency coupling of nonlinear ultrasonic modulation method reflects the generation and development of microcracks in thermal damaged concrete. The peak value of DI increases sensitively with the increasing of water-cement ratio, fine-coarse aggregate ratio, and the heating temperature. Meanwhile, the statistical relationships of the peak value of DI with the residual strength and the area ratio of microcracks in thermal damaged concrete are established respectively.","PeriodicalId":180083,"journal":{"name":"Proceedings of the 13th International Workshop on Structural Health Monitoring","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133256390","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 behavior and performance of engineered structures has become increasingly desirable due to the value such information offers for occupant safety and structural maintenance. Vibration data collected from accelerometers has proven to be an effective tool to perform this type of monitoring. While some monitoring activities can occur autonomously, it is often necessary for humans to interact with the data to discern the need for additional evaluation. In large structures or those with a dense sensor deployment, continuously collected vibration data can quickly grow to massive scales. Consequently, the evaluation of structural performance is often limited by the ability of a system to efficiently process and present large volumes of data. To overcome this challenge, this paper presents a framework to process, store, and visualize data using open-source distributed computing technologies. The framework utilizes a publish-subscribe messaging queue deployed across multiple partitions to consume data in parallel, improving the rate of ingestion. Ingested data is stored in a structured format using a NoSQL database that provides high availability, scalability, and performance. The stored data acts as the source for webbased visualization. This setup provides a high degree of adaptability, allowing meaningful visualizations to be implemented for various forms of smart infrastructure monitoring tasks. The capabilities of the resultant human-infrastructure interface are demonstrated using Goodwin Hall, a five-story building instrumented with 225 hard-wired accelerometers. This case study showcases visualizations that enable users to perform real-time assessment of frequency domain features and efficiently identify notable excitation events during the building's history.
{"title":"A HIGH-VOLUME PROCESSING FRAMEWORK FOR HUMAN-STRUCTURE INTERFACES IN SMART INFRASTRUCTURE","authors":"Natasha Vipond, Abhinav Kumar, Zhiwu Xie, Rodrigo Sarlo","doi":"10.12783/shm2021/36252","DOIUrl":"https://doi.org/10.12783/shm2021/36252","url":null,"abstract":"Monitoring the behavior and performance of engineered structures has become increasingly desirable due to the value such information offers for occupant safety and structural maintenance. Vibration data collected from accelerometers has proven to be an effective tool to perform this type of monitoring. While some monitoring activities can occur autonomously, it is often necessary for humans to interact with the data to discern the need for additional evaluation. In large structures or those with a dense sensor deployment, continuously collected vibration data can quickly grow to massive scales. Consequently, the evaluation of structural performance is often limited by the ability of a system to efficiently process and present large volumes of data. To overcome this challenge, this paper presents a framework to process, store, and visualize data using open-source distributed computing technologies. The framework utilizes a publish-subscribe messaging queue deployed across multiple partitions to consume data in parallel, improving the rate of ingestion. Ingested data is stored in a structured format using a NoSQL database that provides high availability, scalability, and performance. The stored data acts as the source for webbased visualization. This setup provides a high degree of adaptability, allowing meaningful visualizations to be implemented for various forms of smart infrastructure monitoring tasks. The capabilities of the resultant human-infrastructure interface are demonstrated using Goodwin Hall, a five-story building instrumented with 225 hard-wired accelerometers. This case study showcases visualizations that enable users to perform real-time assessment of frequency domain features and efficiently identify notable excitation events during the building's history.","PeriodicalId":180083,"journal":{"name":"Proceedings of the 13th International Workshop on Structural Health Monitoring","volume":" 16","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132187480","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}
In the aftermath of earthquakes, structures can become unsafe and hazardous for humans to safely reside. Automated methods that detect structural damage can be invaluable for rapid inspections and faster recovery times. Deep neural networks (DNNs) have proven to be an effective means to classify damaged areas in images of structures but have limited generalizability due to the lack of large and diverse annotated datasets (e.g., variations in building properties like size, shape, color). Given a dataset of paired images of damaged and undamaged structures supervised deep learning methods could be employed, but such paired correspondences of images required for training are exceedingly difficult to acquire. Obtaining a variety of undamaged images, and a smaller set of damaged images is more viable. We present a novel application of deep learning for unpaired image-to-image translation between undamaged and damaged structures as a means of data augmentation to combat the lack of diverse data. Unpaired image-to-image translation is achieved using Cycle Consistent Adversarial Network (CCAN) architectures, which have the capability to translate images while retaining the geometric structure of an image. We explore the capability of the original CCAN architecture, and propose a new architecture for unpaired image-to-image translation (termed Eigen Integrated Generative Adversarial Network or EIGAN) that addresses shortcomings of the original architecture for our application. We create a new unpaired dataset to translate an image between domains of damaged and undamaged structures. The dataset created consists of a set of damaged and undamaged buildings from Mexico City affected by the 2017 Puebla earthquake. Qualitative and quantitative results of the various architectures are presented to better compare the quality of the translated images. A comparison is also done on the performance of DNNs trained to classify damaged structures using generated images. The results demonstrate that targeted image-to-image translation of undamaged to damaged structures is an effective means of data augmentation to improve network performance.
{"title":"IMAGE TO IMAGE TRANSLATION OF STRUCTURAL DAMAGE USING GENERATIVE ADVERSARIAL NETWORKS","authors":"Subin Varghese, Rebecca Wang, Vedhus Hoskere","doi":"10.12783/shm2021/36307","DOIUrl":"https://doi.org/10.12783/shm2021/36307","url":null,"abstract":"In the aftermath of earthquakes, structures can become unsafe and hazardous for humans to safely reside. Automated methods that detect structural damage can be invaluable for rapid inspections and faster recovery times. Deep neural networks (DNNs) have proven to be an effective means to classify damaged areas in images of structures but have limited generalizability due to the lack of large and diverse annotated datasets (e.g., variations in building properties like size, shape, color). Given a dataset of paired images of damaged and undamaged structures supervised deep learning methods could be employed, but such paired correspondences of images required for training are exceedingly difficult to acquire. Obtaining a variety of undamaged images, and a smaller set of damaged images is more viable. We present a novel application of deep learning for unpaired image-to-image translation between undamaged and damaged structures as a means of data augmentation to combat the lack of diverse data. Unpaired image-to-image translation is achieved using Cycle Consistent Adversarial Network (CCAN) architectures, which have the capability to translate images while retaining the geometric structure of an image. We explore the capability of the original CCAN architecture, and propose a new architecture for unpaired image-to-image translation (termed Eigen Integrated Generative Adversarial Network or EIGAN) that addresses shortcomings of the original architecture for our application. We create a new unpaired dataset to translate an image between domains of damaged and undamaged structures. The dataset created consists of a set of damaged and undamaged buildings from Mexico City affected by the 2017 Puebla earthquake. Qualitative and quantitative results of the various architectures are presented to better compare the quality of the translated images. A comparison is also done on the performance of DNNs trained to classify damaged structures using generated images. The results demonstrate that targeted image-to-image translation of undamaged to damaged structures is an effective means of data augmentation to improve network performance.","PeriodicalId":180083,"journal":{"name":"Proceedings of the 13th International Workshop on Structural Health Monitoring","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134377205","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}
Structural health monitoring (SHM) of additively manufactured polymer parts is challenging due to the very strong attenuation of the surface waves. To excite the part surface at a very wide frequency band in a very short time, Multiple Width Pulse Excitation (MWPE) signal was introduced. MPWE was used to excite the surface of the structure for the implementation of the Surface Response to Excitation (SuRE) method. A cross-shaped polymer part was fabricated additively for the identification of the hidden geometry of the infill. The part had four extensions with identical geometry but different internal designs. Two of the extensions had cross infills and the other two had square infills. For each type of infill, one extension had 1 mm and the other extension had 2 mm thick skin. The part was excited at the middle with WMPE excitation and the dynamic response was monitored at the end of each extension. The Short-Time Fast Fourier Transform (STFFT) was used for the analysis of the signal in the time-frequency domain. The two dimentional sum of the squares of the differences (2DSSD) was used for the classification of the signal. Compressive force and type of infill was identified accurately for all the test cases.
由于表面波的衰减非常强，增材制造聚合物部件的结构健康监测(SHM)具有挑战性。为了在极短的时间内对零件表面进行极宽频带的激励，引入了多宽脉冲激励(MWPE)信号。利用MPWE对结构表面进行激励，实现表面激励响应(surface Response to Excitation, SuRE)方法。为了识别填充物的隐藏几何形状，采用增材制造了十字形聚合物零件。该部件有四个具有相同几何形状但内部设计不同的扩展部分。其中两个扩展部分是交叉填充，另外两个是方形填充。对于每种类型的填充物，一个延伸有1毫米厚，另一个延伸有2毫米厚的皮肤。采用WMPE励磁法对中间部分进行激励，并在每次延伸结束时监测其动态响应。采用短时快速傅立叶变换(STFFT)对信号进行时频分析。采用二维差分平方和(2DSSD)对信号进行分类。所有试验用例的压缩力和填充物类型都得到了准确的识别。
{"title":"NEW EXCITATION (MULTIPLE WIDTH PULSE EXCITATION (MWPE)) METHOD FOR SHM SYSTEMS—PART 1: VISUALIZATION OF TIME- FREQUENCY DOMAIN CHARACTERISTICS","authors":"I. Tansel, Alireza Modir","doi":"10.12783/shm2021/36341","DOIUrl":"https://doi.org/10.12783/shm2021/36341","url":null,"abstract":"Structural health monitoring (SHM) of additively manufactured polymer parts is challenging due to the very strong attenuation of the surface waves. To excite the part surface at a very wide frequency band in a very short time, Multiple Width Pulse Excitation (MWPE) signal was introduced. MPWE was used to excite the surface of the structure for the implementation of the Surface Response to Excitation (SuRE) method. A cross-shaped polymer part was fabricated additively for the identification of the hidden geometry of the infill. The part had four extensions with identical geometry but different internal designs. Two of the extensions had cross infills and the other two had square infills. For each type of infill, one extension had 1 mm and the other extension had 2 mm thick skin. The part was excited at the middle with WMPE excitation and the dynamic response was monitored at the end of each extension. The Short-Time Fast Fourier Transform (STFFT) was used for the analysis of the signal in the time-frequency domain. The two dimentional sum of the squares of the differences (2DSSD) was used for the classification of the signal. Compressive force and type of infill was identified accurately for all the test cases.","PeriodicalId":180083,"journal":{"name":"Proceedings of the 13th International Workshop on Structural Health Monitoring","volume":"194 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132280352","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}
3D remote sensing technologies have improved dramatically over the past five years and methods such as laser scanning and photogrammetry are now capable of reliably resolving geometric details on the order of one millimeter or less. This has significant impacts for the structural health monitoring community, as it has expanded the range of mechanics-driven problems that these methods can be employed on. In this work, we explore how 3D geometric measurements extracted from photogrammetric point clouds can be leveraged for structural analysis and measurement of structural deformations without physically contacting the target structure. Here we present a non-destructive evaluation technique for extracting and quantifying structural deformations as applied to a load test on a highway bridge in Delaware. The challenging nature of 3D point cloud data means that statistical methods must be employed to adequately evaluate the deformation field of the bridge. Overall, the results show a direct pathway from 3D imaging to fundamental mechanical analysis with measurements that capture the true deformation values typically within one standard deviation. These results are promising given that the mid-span deformation of the bridge for the given load test is on the scale of only a few millimeters. Future work for this method will also investigate using these results for updating finite element models.
{"title":"FULL-SCALE DEFORMATION FIELD MEASUREMENTS VIA PHOTOGRAMMETRIC REMOTE SENSING","authors":"W. Graves, D. Lattanzi","doi":"10.12783/shm2021/36298","DOIUrl":"https://doi.org/10.12783/shm2021/36298","url":null,"abstract":"3D remote sensing technologies have improved dramatically over the past five years and methods such as laser scanning and photogrammetry are now capable of reliably resolving geometric details on the order of one millimeter or less. This has significant impacts for the structural health monitoring community, as it has expanded the range of mechanics-driven problems that these methods can be employed on. In this work, we explore how 3D geometric measurements extracted from photogrammetric point clouds can be leveraged for structural analysis and measurement of structural deformations without physically contacting the target structure. Here we present a non-destructive evaluation technique for extracting and quantifying structural deformations as applied to a load test on a highway bridge in Delaware. The challenging nature of 3D point cloud data means that statistical methods must be employed to adequately evaluate the deformation field of the bridge. Overall, the results show a direct pathway from 3D imaging to fundamental mechanical analysis with measurements that capture the true deformation values typically within one standard deviation. These results are promising given that the mid-span deformation of the bridge for the given load test is on the scale of only a few millimeters. Future work for this method will also investigate using these results for updating finite element models.","PeriodicalId":180083,"journal":{"name":"Proceedings of the 13th International Workshop on Structural Health Monitoring","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133762356","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}
Gaussian process regression is a powerful method for predicting states associated with uncertainty. A common application field is to predict damage states of structural systems. Recently, Gaussian processes became very popular as they deliver credible intervals for the predicted states. However, one major disadvantage of Gaussian processes is that they assume a normal distribution. This is not justified when the relevant variables can only assume positive values, such as crack lengths or damage states. This paper presents a way to bypass this problem by using warped Gaussian processes: We (1) transform the data with a warping function, (2) apply Gaussian process regression in the latent space, and (3) transform the results back by using the inverse of the warping function. The method is applied to a crack growth example. The paper shows how to integrate prior knowledge into warped Gaussian processes in order to increase prediction accuracy and that warped Gaussian processes lead to better and more plausible results.
{"title":"WARPED GAUSSIAN PROCESSES FOR PROGNOSTIC HEALTH MONITORING","authors":"Simon Pfingstl, Christian Braun, M. Zimmermann","doi":"10.12783/shm2021/36358","DOIUrl":"https://doi.org/10.12783/shm2021/36358","url":null,"abstract":"Gaussian process regression is a powerful method for predicting states associated with uncertainty. A common application field is to predict damage states of structural systems. Recently, Gaussian processes became very popular as they deliver credible intervals for the predicted states. However, one major disadvantage of Gaussian processes is that they assume a normal distribution. This is not justified when the relevant variables can only assume positive values, such as crack lengths or damage states. This paper presents a way to bypass this problem by using warped Gaussian processes: We (1) transform the data with a warping function, (2) apply Gaussian process regression in the latent space, and (3) transform the results back by using the inverse of the warping function. The method is applied to a crack growth example. The paper shows how to integrate prior knowledge into warped Gaussian processes in order to increase prediction accuracy and that warped Gaussian processes lead to better and more plausible results.","PeriodicalId":180083,"journal":{"name":"Proceedings of the 13th International Workshop on Structural Health Monitoring","volume":"103 3","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133007883","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：