Data-driven methods in structural health monitoring (SHM) is gaining popularity due to recent technological advancements in sensors, as well as high-speed internet and cloud-based computation. Since the introduction of deep learning (DL) in civil engineering, particularly in SHM, this emerging and promising tool has attracted significant attention among researchers. The main goal of this paper is to review the latest publications in SHM using emerging DL-based methods and provide readers with an overall understanding of various SHM applications. After a brief introduction, an overview of various DL methods (e.g., deep neural networks, transfer learning, etc.) is presented. The procedure and application of vibration-based, vision-based monitoring, along with some of the recent technologies used for SHM, such as sensors, unmanned aerial vehicles (UAVs), etc. are discussed. The review concludes with prospects and potential limitations of DL-based methods in SHM applications.

This study introduces a novel convolutional neural network (CNN)‐based approach for structural health monitoring (SHM) that exploits a form of measured compressed response data through transfer learning (TL)‐based techniques. The implementation of the proposed methodology allows damage identification and localization within a realistic large‐scale system. To validate the proposed method, first, a well‐known benchmark model is numerically simulated. Using acceleration response histories, as well as compressed response data in terms of discrete histograms, CNN models are trained, and the robustness of the CNN architectures is evaluated. Finally, pretrained CNNs are fine‐tuned to be adaptable for three‐parameter, extremely compressed response data, based on the response mean, standard deviation, and a scale factor. The performance of each CNN implementation is assessed using training accuracy histories as well as confusion matrices, along with other performance metrics. In addition to the numerical study, the performance of the proposed method is demonstrated using experimental vibration response data for verification and validation. The results indicate that deep TL can be implemented effectively for SHM of similar structural systems with different types of sensors.

Dynamic characteristics of cable-stayed bridges are widely accepted as valuable indicators to determine their performance in structural health monitoring (SHM). Although research has been extensively conducted in this area, such vibration-based physics methods still face great challenges in improving the effectiveness of damage identification from complex large-scale systems, particularly when other factors, including operational and environmental conditions, may cause high interference to the vibration response. Data-intensive machine learning techniques have been gaining attention due to their robustness for data classification. In this study, a framework was developed for data-driven structural diagnosis and damage detection using a support vector machine (SVM) integrated with enhanced feature extraction techniques for rapid condition assessment for large-scale cable-stayed bridges. The wavelet transform, Hilbert-Huang transform (HHT), and Teager-Huang transform (THT) were selected as three representative feature extraction methods. A kernel function-based SVM was used to facilitate the identification of damaged and undamaged cases. Numerical simulation was conducted to verify the effectiveness and accuracy of the proposed methods applied to a cable-stayed bridge. Results showed that the wavelet time-frequency analysis is more robust to noise than the HHT and THT, whereas the latter two transforms are more sensitive to capture damage/defects. Moreover, for regular signal data, the THT, due to the high time resolution, had the highest concentration and thus is the most sensitive compared with the other two methods. Parameters of interest, including impacts of damage level, damage location, sensor locations, and moving vehicle loading, are extensively discussed. All cases reveal that data-driven approaches could effectively map damage features over and under undamaged cases, dramatically enhancing the effectiveness and accuracy of data classification, which will greatly benefit in situ cable-stayed bridge assessment and management.