Search

[Display omitted] • Cracks are monitored using distributed fiber optic sensor (DFOS) and deep learning. • A modified You Only Look Once (YOLO) model adequately interprets DFOS data. • Transfer learning is incorporated to improve the accuracy of the deep learning model. • The robustness of the proposed approach is evaluated in different test scenarios. • The mAP@0.5 of detecting spatially-distributed cracks reaches 0.968. Distributed fiber optic sensors (DFOSs) offer unique capabilities for crack monitoring via measuring strain distributions. However, manually interpreting strain distributions is labor-intensive and time-consuming. To address this challenge, this paper presents a deep learning approach for real-time automatic interpretation of strain distributions, aiming at monitoring spatially-distributed cracks. The proposed approach encompasses three key innovations. First, deep learning-based methods are developed to facilitate automatic detection and localization of spatially-distributed cracks. Second, transfer learning is incorporated to overcome the data scarcity issue in training deep learning models. This ensures robust performance even with limited data. Third, a split-and-merge method is developed, enhancing the accuracy of multi-crack detection. To evaluate the performance of the approach, experimental data from various cases were considered. The results demonstrate a mean average precision (mAP) of 0.968 for crack detection. The processing time for a set of DFOS data, containing 10,000 measurement points, was less than 0.05 s. [ABSTRACT FROM AUTHOR]

Distributed fiber optic sensors have exhibited superior capabilities in monitoring cracks in engineering structures through measuring detailed strain distributions. However, manually interpreting the measurements from long distributed sensors deployed in large-scale structures is time-consuming, labor-intensive, and subject to human errors. This paper proposes to automate the identification, localization, quantification, and visualization of cracks through intelligent interpretation of strain distributions measured from distributed fiber optic sensors based on machine learning. Based on these intelligent capabilities, a live digital twin model based on building information modeling is developed to visualize cracks. The digital twin model is updatable with real-time measurements from strain distributions from distributed fiber optic sensors. The proposed approach is evaluated via laboratory testing of a concrete beam. The results show that the proposed approach achieves high accuracy in interpretation of sensor data for crack monitoring. This research advances the capabilities of structural health monitoring using distributed fiber optic sensors. [Display omitted] • Automatic interpretation of strain distributions from distributed fiber optic sensors is enabled. • Machine learning approaches are presented to identify, locate, quantify, and visualize cracks. • A digital twin model is developed to visualize cracks and updated by real-time measurement. • Unsupervised and supervised learning approaches achieve high accuracy and high efficiency. [ABSTRACT FROM AUTHOR]

[Display omitted] Recent advances in robotics and machine vision enable automated measurement of vibration for civil engineering structures. Various types of robotic platforms such as drones and land robots instrumented with advanced navigation and data acquisition systems were developed. Particularly, various cameras have been deployed on robotic platforms to capture videos of structures, and machine vision approaches were developed to extract structural vibration from captured videos. This forward-looking review discusses the robot-based hardware, the machine vision approaches, and the capabilities of automated measurement of structural vibration. Emphases are placed on the clarifying the capabilities, technical limitations, promising solutions, and practical performance of robot-based approaches in field measurements that involve low vibration amplitudes and long-term monitoring requirements. Challenges and opportunities are discussed, and recommendations are provided to facilitate practical applications. This review will encourage the development and applications of robots for automated condition assessment of civil engineering structures. [ABSTRACT FROM AUTHOR]

Measuring distance is essential for health monitoring and condition assessment of civil engineering structures. This paper reviews recent advances in remote sensing technologies for measuring distance based on electromagnetic waves. Specifically, four families of technologies are reviewed, which are the Global Navigation Satellite Systems, microwave radars, laser-based methods, and vision-based methods. The reviewed content covers the measurement principles, signal processing methods, state-of-the-art applications, and key performance metrics. The investigated performance includes the measurement accuracy, sampling frequency, operating distance, robustness to the environment, and compatibility with autonomous platforms. Existing inconsistent viewpoints concerning the performance are discussed. Based on the features of different technologies, a decision tree is presented to facilitate selection of appropriate methods for intended applications. This research is expected to promote development and applications of remote sensing technologies for facilitating condition assessment of engineering structures. [ABSTRACT FROM AUTHOR]

• A method is proposed to measure crack widths using a single distributed fiber optic sensor. • The measurement accuracy of crack width using the distributed sensor is as fine as 5.6 µm. • The widths of multiple cracks can be measured using the same distributed sensor. • Influences of spatial resolution, coating thickness, and crack spacing are investigated. • The tradeoff of the maximum measurable crack width and the spacing is quantified. This study develops a method to measure crack width using a distributed fiber optic sensor. To this end, a customized specimen was designed to manipulate cracks under precise displacement control. An optical fiber was attached on the specimen and served as a distributed sensor which measured strain distributions along the fiber based on optical frequency domain reflectometry. An algorithm was presented to analyze crack width based on the measured strain distribution. Compared with the crack width measured by an extensometer, the crack width measured by the distributed sensor had an accuracy of 5.6 µm. Parametric studies were conducted and revealed that the coating thickness of optical fiber, the spatial resolution of strain distribution, and the spacing of cracks had significant effects on the measurement of cracks using the distributed sensor. This study is expected to enhance the capability of detecting, locating, and quantifying cracks by using distributed fiber optic sensors. [ABSTRACT FROM AUTHOR]

[Display omitted] • Pipeline corrosion is monitored using distributed fiber optic sensors and machine learning. • Distributed fiber optic sensor data are automatically interpreted for monitoring corrosion. • Bayesian optimization is used to enhance the accuracy of the machine learning model. • Mass loss due to steel corrosion in pipelines is quantified. • The robustness of the proposed approach is investigated in various scenarios. Distributed fiber optic sensor (DFOS) offers unique capabilities of monitoring corrosion for long pipelines. However, manually interpreting DFOS data is labor-intensive and time-consuming. To address this challenge, this paper presents a machine learning approach for real-time automatic interpretation of DFOS data used to monitor both uniform and non-uniform corrosion in pipeline. A machine learning model is developed to automatically detect corrosion based on DFOS data, and a corrosion quantification method is developed based on the output of the machine learning model. The proposed approaches are evaluated using laboratory experiments in terms of accuracy and robustness to pipeline diameter, spatial resolution of DFOS, type of fiber optic cable, and sensor installation methods. The results show that the F1 score for corrosion detection and the R2 value for corrosion quantification are 0.986 and 0.953, respectively. This research will facilitate pipeline corrosion monitoring by enabling automatic distributed sensor data interpretation. [ABSTRACT FROM AUTHOR]

• The method is developed based on long-term recurrent convolutional network. • The network model is trained using videos captured using a cellphone camera. • The proposed method can achieve real-time (<1 s) plastic viscosity assessment. • The effects and mechanisms of plastic viscosity on flexural properties are studied. Plastic viscosity is a key property of ultra-high-performance concrete (UHPC) and must be controlled during the mixing to achieve desired fresh and hardened properties. This paper presents a video recognition technology for real-time assessment of plastic viscosity using a video captured by a camera during the mixing of UHPC. A long-term recurrent convolutional network is proposed to extract spatial and temporal features of flowing UHPC from the video and correlate the features with plastic viscosity measured from a rheometer, thus enabling assessment of plastic viscosity using videos. This research also investigates the effects of plastic viscosity on the fiber dispersion and orientation, air content, and flexural properties of UHPC. The results show that the plastic viscosity significantly influences fiber distribution and air void content, thus affecting the flexural properties of UHPC. The presented method enables real-time assessment of plastic viscosity for control of flexural properties and air void content of UHPC. This study will greatly facilitate quality control for production of UHPC. [ABSTRACT FROM AUTHOR]

• Distributed fiber optic sensors are used to detect buckling for thin-walled structures. • An effective and practical method is developed for shape reconstruction using strain fields. • Strain fields due to buckling are measured by fully distributed sensors with sub-mm resolution. • The proposed shape reconstruction method is compared with a computer-vision method. • Effects of spatial resolution and sensor deployment on shape reconstruction are discussed. This paper proposes to detect buckling and reconstruct three-dimensional deformations using distributed fiber optic sensors. The distributed sensors measured strain distributions with sub-millimeter resolutions based on optical frequency domain reflectometry in real time. Buckling was detected from high-resolution strain distributions measured from specimens. An effective and practical shape reconstruction approach was developed to derive nonsymmetrical deformations based on the strain distributions. The reconstructed shape was validated using a computer vision method that measured point cloud from the specimens. A parametric study was conducted to investigate the effects of the key sensing parameters such as the spatial resolution and sensor deployment scheme on the performance of the shape reconstruction approach, and used to optimize the resolution and deployment of distributed sensors. This research will advance the capability of buckling detection and shape reconstruction through distributed sensing for engineering structures under complex loading conditions. [ABSTRACT FROM AUTHOR]