Explore the vast range of titles available.

Large-volume hydraulic concrete structures, such as concrete dams, often suffer from damage due to the influence of alternating loads and material aging during the service process. The occurrence and further expansion of cracks will affect the integrity, impermeability, and durability of the dam concrete. Therefore, monitoring the changing status of cracks in hydraulic concrete structures is very important for the health service of hydraulic engineering. This study combines computer vision and artificial intelligence methods to propose an automatic damage detection and diagnosis method for hydraulic structures. Specifically, to improve the crack feature extraction effect, the Xception backbone network, which has fewer parameters than the ResNet backbone network, is adopted. With the aim of addressing the problem of premature loss of image detail information and small target information of tiny cracks in hydraulic concrete structures, an adaptive attention mechanism image semantic segmentation algorithm based on Deeplab V3+ network architecture is proposed. Crack images collected from concrete structures of different types of hydraulic structures were used to develop crack datasets. The experimental results show that the proposed method can realize high-precision crack identification, and the identification results have been obtained in the test set, achieving 90.537% Intersection over Union (IOU), 91.227% Precision, 91.301% Recall, and 91.264% F1_score. In addition, the proposed method has been verified on different types of cracks in actual hydraulic concrete structures, further illustrating the effectiveness of the method.

Remotely sensed data can provide the basis for timely and efficient building damage maps that are of fundamental importance to support the response activities following disaster events. However, the generation of these maps continues to be mainly based on the manual extraction of relevant information in operational frameworks. Considering the identification of visible structural damages caused by earthquakes and explosions, several recent works have shown that Convolutional Neural Networks (CNN) outperform traditional methods. However, the limited availability of publicly available image datasets depicting structural disaster damages, and the wide variety of sensors and spatial resolution used for these acquisitions (from space, aerial and UAV platforms), have limited the clarity of how these networks can effectively serve First Responder needs and emergency mapping service requirements. In this paper, an advanced CNN for visible structural damage detection is tested to shed some light on what deep learning networks can currently deliver, and its adoption in realistic operational conditions after earthquakes and explosions is critically discussed. The heterogeneous and large datasets collected by the authors covering different locations, spatial resolutions and platforms were used to assess the network performances in terms of transfer learning with specific regard to geographical transferability of the trained network to imagery acquired in different locations. The computational time needed to deliver these maps is also assessed. Results show that quality metrics are influenced by the composition of training samples used in the network. To promote their wider use, three pre-trained networks—optimized for satellite, airborne and UAV image spatial resolutions and viewing angles—are made freely available to the scientific community.

Summary To avoid false alarms for vibration‐based structural damage detection methods, temperature effects on damage‐sensitive features should be eliminated. In this paper, a novel two‐step damage identification method combining a multilayer neural network and novelty detection is developed to differentiate the changes in natural frequencies (one of the most commonly used damage features that can be obtained reliably and relatively easily) due to damage from those induced by temperature variations. In the first step, a multilayer artificial neural network, which resembles an auto‐associative neural network but uses temperature variables in addition to the frequencies as the inputs, is explored to identify patterns in frequencies of undamaged structures under varying temperatures. Euclidean distance is then utilized as a novelty index to quantify the discordancy between patterns in undamaged cases and candidate cases. Numerical studies using a simply supported beam and finite element models based on an experimental grid structure, which simulate different levels of stiffness reductions under varying temperature conditions, are used to verify the detectability and robustness of the proposed approach. It is shown that the incorporation of the proposed artificial neural network with novelty detection enables one to robustly distinguish damage occurrence and severity regardless of temperature variations and noise perturbations. Using an unsupervised learning scheme, the proposed approach transforms a multivariate analysis using modal frequencies and temperature data into a straightforward univariate discordancy test using the novelty index. Given these competitive advantages, this approach is very attractive for the development of an automated continuous monitoring system in practical applications.

Structural damage identification based on finite element (FE) model updating has been a research direction of increasing interest over the last decade in the mechanical, civil, aerospace, etc., engineering fields. Various studies have addressed direct, sensitivity-based, probabilistic, statistical, and iterative methods for updating FE models for structural damage identification. In contrast, evolutionary algorithms (EAs) are a type of modern method for FE model updating. Structural damage identification using FE model updating by evolutionary algorithms is an active research focus in progress but lacking a comprehensive survey. In this situation, this study aims to present a review of critical aspects of structural damage identification using evolutionary algorithm-based FE model updating. First, a theoretical background including the structural damage detection problem and the various types of FE model updating approaches is illustrated. Second, the various residuals between dynamic characteristics from FE model and the corresponding physical model, used for constructing the objective function for tracking damage, are summarized. Third, concerns regarding the selection of parameters for FE model updating are investigated. Fourth, the use of evolutionary algorithms to update FE models for damage detection is examined. Fifth, a case study comparing the applications of two single-objective EAs and one multi-objective EA for FE model updating-based damage detection is presented. Finally, possible research directions for utilizing evolutionary algorithm-based FE model updating to solve damage detection problems are recommended. This study should help researchers find crucial points for further exploring theories, methods, and technologies of evolutionary algorithm-based FE model updating for structural damage detection.

Automatic detection of corrosion and associated damages to civil infrastructures such as bridges, buildings, and roads, from aerial images captured by an Unmanned Aerial Vehicle (UAV), helps one to overcome the challenges and shortcomings (objectivity and reliability) associated with the manual inspection methods. Deep learning methods have been widely reported in the literature for civil infrastructure corrosion detection. Among them, convolutional neural networks (CNNs) display promising applicability for the automatic detection of image features less affected by image noises. Therefore, in the current study, we propose a modified version of deep hierarchical CNN architecture, based on 16 convolution layers and cycle generative adversarial network (CycleGAN), to predict pixel-wise segmentation in an end-to-end manner using the images of Bolte Bridge and sky rail areas in Victoria (Melbourne). The convolutedly designed model network proposed in the study is based on learning and aggregation of multi-scale and multilevel features while moving from the low convolutional layers to the high-level layers, thus reducing the consistency loss in images due to the inclusion of CycleGAN. The standard approaches only use the last convolutional layer, but our proposed architecture differs from these approaches and uses multiple layers. Moreover, we have used guided filtering and Conditional Random Fields (CRFs) methods to refine the prediction results. Additionally, the effectiveness of the proposed architecture was assessed using benchmarking data of 600 images of civil infrastructure. Overall, the results show that the deep hierarchical CNN architecture based on 16 convolution layers produced advanced performances when evaluated for different methods, including the baseline, PSPNet, DeepLab, and SegNet. Overall, the extended method displayed the Global Accuracy (GA); Class Average Accuracy (CAC); mean Intersection Of the Union (IOU); Precision (P); Recall (R); and F-score values of 0.989, 0.931, 0.878, 0.849, 0.818 and 0.833, respectively.

Rapid detecting damages/defeats in the large-scale civil engineering structures, assessing their conditions and timely decision making are crucial to ensure their health and ultimately enhance the level of public safety. Advanced sensor network techniques recently allow collecting large amounts of data for structural health monitoring and damage detection, while how to effectively interpret these complex sensor data to technical information posts many challenges. This paper presents three optimization-algorithm based support vector machines for damage detection. The optimization algorithms, including grid-search, partial swarm optimization and genetic algorithm, are used to optimize the penalty parameters and Gaussian kernel function parameters. Two types of feature extraction methods in terms of time-series data are selected to capture effective damage characteristics. A benchmark experimental data with the 17 different scenarios in the literature were used for verifying the proposed data-driven methods. Numerical results revealed that all three optimized machine learning methods exhibited significantly improvement in sensitivity, accuracy and effectiveness over conventional methods. The genetic algorithm based SVM had a better prediction than other methods. Two different feature methods used in this study also demonstrated the appropriate features are crucial to improve the sensitivity in detecting damage and assessing structural health conditions. The findings of this study are expected to help engineers to process big data and effectively detect the damage/defects, and thus enable them to make timely decision for supporting civil infrastructure management practices.

While many structural damage detection methods have been developed in recent decades, few data‐driven methods in unsupervised learning mode have been developed to solve the practical difficulties in data acquisition for civil infrastructures in different scenarios. To address such a challenge, this article proposes a number of improved unsupervised novelty detection methods and conducts extensive comparative studies on a laboratory scale steel bridge to examine their performances of damage detection. The key concept behind unsupervised novelty detection in this article is that only normal data from undamaged/baseline structural scenarios are required to train statistical models with these methods. Then, these trained models are used to identify abnormal testing data from damaged scenarios. To detect structural damage in the form of loosening bolts in the steel bridge, four machine‐learning methods (i.e., K‐nearest neighbors method, Gaussian mixture models, one‐class support vector machines, density peaks‐based fast clustering method) and one deep learning method using a deep auto‐encoder are selected. Meanwhile, some modifications and improvements are made to enable these methods to detect structural damage in unsupervised novelty detection mode. In their comparative studies, the advantages and disadvantages of these methods are analyzed based on their results of structural damage detection. Recently, deep learning‐based damage detection is a very hot topic. This article conducted extensive comparative studies using state‐of‐the‐art methods of deep learning‐based damage detection methods to figure out the pros and cons of each method.

Structural damage detection is of very importance to improve reliability and safety of civil structures. A novel sensor data-driven structural damage detection method is proposed in this paper by combining continuous wavelet transform (CWT) with deep convolutional neural network (DCNN). In this method, time-frequency images are obtained by CWT from original one-dimensional sensor signals. And, DCNN is designed to mine structural damage features from the time-frequency images and distinguish different structural damage condition. The proposed method is carried out on three-story building structure dataset and steel frame dataset. The experimental results show that the proposed method has the high accuracy and robustness of the damage detection compared with other existing machine learning methods.

Remote sensing is an effective method of evaluating building damage after a large-scale natural disaster, such as an earthquake or a typhoon. In recent years, with the development of computer vision technology, deep learning algorithms have been used for damage assessment from aerial images. In April 2016, a series of earthquakes hit the Kyushu region, Japan, and caused severe damage in the Kumamoto and Oita Prefectures. Numerous buildings collapsed because of the strong and continuous shaking. In this study, a deep learning model called Mask R-CNN was modified to extract residential buildings and estimate their damage levels from post-event aerial images. Our Mask R-CNN model employs an improved feature pyramid network and online hard example mining. Furthermore, a non-maximum suppression algorithm across multiple classes was also applied to improve prediction. The aerial images captured on 29 April 2016 (two weeks after the main shock) in Mashiki Town, Kumamoto Prefecture, were used as the training and test sets. Compared with the field survey results, our model achieved approximately 95% accuracy for building extraction and over 92% accuracy for the detection of severely damaged buildings. The overall classification accuracy for the four damage classes was approximately 88%, demonstrating acceptable performance.

The response of the San Pietro monumental bell-tower located in Perugia, Italy, to the 2016 Central Italy seismic sequence is investigated, taking advantage of the availability of field data recorded by a vibration-based SHM system installed in December 2014 to detect earthquake-induced damages. The tower is located about 85 km in the NW direction from the epicenter of the first major shock of the sequence, the Accumoli Mw6.0 earthquake of August 24th, resulting in a small local PGA of about 30 cm/s2, whereby near-field PGA was measured as 915.97 cm/s2 (E–W component) and 445.59 cm/s2 (N–S component). Similar PGA values also characterized the two other major shocks of the sequence (Ussita Mw5.9 and Norcia Mw6.5 earthquakes of October 26th and 30th, respectively). Despite the relatively low intensity of such earthquakes in Perugia, the analysis of long-term monitoring data clearly highlights that small permanent changes in the structural behavior of the bell-tower have occurred after the earthquakes, with decreases in all identified natural frequencies. Such natural frequency decays are fully consistent with what predicted by non-linear finite element simulations and, in particular, with the development of microcracks at the base of the columns of the belfry. Microcracks in these regions, and in the rest of tower, are however hardly distinguishable from pre-existing ones and from the physiological cracking of a masonry structure, what validates the effectiveness of the SHM system in detecting earthquake-induced damage at a stage where this is not yet detectable by visual inspections.