Bridge Damage Detection Research Articles

Enhancing bridge damage detection with Mamba-Enhanced HRNet for semantic segmentation.

With the acceleration of urbanization, bridges, as crucial infrastructure, their structural health and stability are paramount to public safety. This paper proposes Mamba-Enhanced HRNet for bridge damage detection. Mamba-Enhanced HRNet integrates the advantages of HRNet's multi-resolution parallel design and VMamba's visual state space model. By replacing the residual convolutional blocks in HRNet with a combination of VSS blocks and convolution, this model enhances the network's capability to capture global contextual information while maintaining computational efficiency. This work builds an extensive dataset with multiple damage kinds and uses Mean Intersection over Union (Mean IoU) as the assessment metric to assess the performance of Mamba-Enhanced HRNet. Experimental results demonstrate that Mamba-Enhanced HRNet achieves significant performance improvements in bridge damage semantic segmentation tasks, with Mean IoU scores of 0.963, outperforming several other semantic segmentation models.

TSCB-Net: transformer-enhanced semantic segmentation of surface damage of concrete bridges

Concrete bridges are susceptible to surface damage such as water-corrosion, spalling, and rebar-corrosion, which can pose serious safety risks, and to detect this is crucial to ensure their safety. Computer vision-based methods to detect surface damages in concrete bridges have seen great progress. However, mainstream methods can only locate surface damage but cannot obtain detailed information, such as the proportion of damaged pixels. To address these challenges, this paper propose TSCB-Net, a semantic segmentation model for surface damage detection in concrete bridges. TSCB-Net employs both Transformers and an encoder-decoder structure, offering a strong alternative to detect surface damage in concrete bridges. Its encoder uses improved baselines with pyramid vision transformer (PVTV2) to extract semantic information and spatial details of concrete bridge images, layer by layer, and to capture long-term dependence between features. The decoder incorporates a context information recovery (CIR) module in the final upsampling stage, using convolutional layers to realize a larger receptive field. This module effectively models context information for semantic segmentation and facilitates multi-scale surface damage detection of concrete bridges. In experiments, TSCB-Net achieves an overall mIoU of 75.84% and an mF1 of 85.54% when the number of parameters is 5.84 M, outperforming state-of-the-art models. TSCB-Net can quickly and accurately detect and identify surface bridge damage, which can help in the early detection of structural problems and the adoption of appropriate repair and protection measures.

Enhanced operational modal analysis and change point detection for vibration-based structural health monitoring of bridges

One of the most promising uses of vibration-based structural health monitoring (VBSHM) in bridge damage detection is the tracking of modes through long-term repeated or continuous operational modal analysis (OMA). Any shifts in modal parameters over time can signal structural damage. However, in real-world applications, noise and environmental uncertainties introduce variability in the data, potentially obscuring damage-related changes. To address this, it is essential to establish and understand the temporal trends and behavior of the estimated modal parameters, enabling accurate interpretation of the engineering data. This paper presents a detailed study focusing on data-driven techniques to improve the OMA results by determining the causes of modal variability and establishing modal models to filter out these known causes of variability. It explores the use of data continuously collected over a period of one month in November 2017 on the Confederation Bridge in eastern Canada. Operational modal analysis is conducted to extract modal frequencies and mode shapes, revealing correlations with environmental and operational factors such as wind, temperature and vehicular traffic. A novel approach using the residuals from regression modal models for damage detection is proposed, utilizing a change point detection algorithm. Results indicate the potential to detect shifts in modal frequencies corresponding to damage scenarios, at lower levels than was previously possible, highlighting the feasibility of using enhanced modal features for sensitive damage identification. Overall, the paper contributes to advancing the understanding of variability in vibration-based structural health monitoring and presents a promising practical technique for improving damage detection results using enhanced operational modal estimates in realistic field applications of a real-world structure.

A portable real-time concrete bridge damage detection system

A drone detection system can achieve real-time inspection of concrete bridge damage. However, it is difficult to deploy existing detection algorithms on a portable system due to their high computational cost. Therefore, a portable damage detection algorithm called CD-YOLOv8 is designed. CD-YOLOv8 is operated on Jetson Xavier NX to detect drone-captured bridge images in real-time. The multi-scale feature adaptive fusion module is used to reduce model parameters and improve detection speed. The spatial context pyramid and deformable convolutional structure are introduced to improve the detection accuracy of tiny concrete bridge damage. Compared with the YOLOv8n, the mAP50 of CD-YOLOv8 is increased by 3.6 %, while the model parameters and FLOPs are decreased by 0.05 M and 0.7 G, respectively. CD-YOLOv8 is also tested on the Jetson Xavier NX with a detection speed of 44 FPS, which is 15.79 % faster than the YOLOv8n. Therefore, CD-YOLOv8 has high accuracy and excellent real-time performance.

Transfer learning for structure damage detection of bridges through dynamic distribution adaptation

In recent years, the rapid increase in the number of bridges not only brings convenience to people but also means that more bridges are at potential safety risks. How to quickly and accurately identify damage that occurs in bridges is a primary problem for society to solve. Machine learning, which can train a classifier for bridge damage detection from existing labeled data, has been widely applied to damage detection of bridge structures. However, the method can have limitations such as the difficulty of collecting sufficient damage data, the heavy workload of the data labeling task, and the failure of the classifier when the bridge structure changes, which limits the application of machine learning in structural damage detection. In order to achieve damage detection of bridges with different spans without target labeling, a dynamic distribution adaptation network (DDAN) containing a feature extraction module and a dynamic distribution adaptation (DDA) module is designed in this paper. DDAN uses the feature extraction module to extract features, and then the DDA module acquires the extracted features to dynamically adapt the probability distribution for calibrating the features in the source and target domains which makes the feature distributions in the source and target domains approximate. In order to verify the feasibility of this paper's method, DDAN is applied to the damage detection of simply supported box girder bridges with different spans through numerical simulation in this paper, where the damage features of a 32 m bridge (source domain) are transferred to a 24 m bridge (target domain), and damage localization and quantification of the target bridge is achieved without damage labels on the target bridge. The results show that DDAN possesses a better classification effect as well as faster convergence speed than the network without domain adaptation and traditional deep domain adaptation networks, and has important application prospects in damage detection of bridges with different spans.

Damage detection for continuous beams by using the tap-scan method

Damage detection for bridges using a passing vehicle has drawn much interest. Recent studies demonstrate that the tap-scan method is an efficient way to detect the change of beam bending stiffness. A quantitative relationship between the change of beam bending stiffness and the acceleration of a vehicle can be evaluated. However, the reported relationship is calibrated for a simply-supported beam, which may be unsuitable for continuous beams due to inherent structural differences. Addressing this, an analytical solution for damage detection for continuous beams using the tap-scan method is presented in this paper. Case studies reveal that damage relationships across different spans of continuous beams show negligible dependence on modal parameters and beam lengths. Consequently, the damage identification within any span of a continuous beam can be accomplished using the same damage relationship calibrated on a simply-supported beam. But it should be noticed that the damage relationship might differ by beam structural forms. The above result holds substantial implications for practical engineering scenarios, notably by diminishing the labor and complexity associated with the calibration of damage relationships for continuous beams.

Railway vibration – fast physics-based models for the prediction of ground vibration and the identification of track damage

The following applications of machine learning/AI for railway vibration will be discussed: 1. The prediction of the wave propagation from a railway line (completely physics based for surface lines, for tunnel lines (partially) physics-based ML for surface lines). 2. The track behaviour for the emission of train-induced ground vibration (physics based for homogeneous soil, ML for layered soil). 3. Track damage detection and quantification from FRF and moving load response. 4. Bridge damage detection and localisation from modal analysis and moving load response. 5. The use of axle-box acceleration for the identification of track/sub-soil condition and bridge resonances. Prediction calculation of railway vibration usually needs time-consuming finite element, boundary element and wavenumber domain calculations. For a user-friendly prediction software, fast calculations are needed. Several time-consuming detailed calculations should be used to develop simpler and fast models. The dynamic stiffness of isolated or un-isolated railway tracks from detailed calculations with a continuous soil is approximated with the simpler Winkler soil. The vehicle-track resonance (P2 resonance) rules the effect of the mitigation measures, and it can also be used for the on-board monitoring of the track and sub-soil condition. For the identification of track damage such as gaps between sleepers, track slabs and layers, detailed models with a continuous soil have been updated to get the best fit to the measured frequency response functions from hammer tests and the deformation pattern from the moving load response. Whereas the track damage can be locally identified, this is more difficult for bridges where the modal analysis gives mainly global information. The influence lines of the inclination for statically passing vehicles (locomotive, truck, compaction ) have been used to localise bridge damage (stiffness variations). The on-board monitoring of rail bridges needs special conditions (regular trains with special speeds) to excite and measure the bridge resonance.

Drive-by bridge damage detection by vehicle-bridge interaction neural operator

Drive-by bridge damage detection using vibrations of vehicles for damage detection of bridges has been studied since it needs no instrumentation on bridges. However, it is still a challenge to be able to extract features that are sensitive to bridge damage and relevant to the responses of the bridge. To address the challenges of drive-by bridge damage detection, machine learning-based approaches have also been actively investigated in recent research. With supervised learning methods, it is almost impossible to obtain training data on the damage state of a bridge. On the other hand, unsupervised learning approaches have been considered, but they only provide changes in relative features rather than physical properties that are useful for estimating the structural performance. To address the above issues, this study aims to propose and investigate the feasibility of drive-by bridge damage detection using a vehicle-bridge interaction neural operator (VINO). In the proposed method, numerical simulations of vehicle-bridge interactions considering the damage field of the bridge are first carried out to generate physically informed training data. The VINO is then constructed using the physically informed training data. However, VINOs are still insufficient to model real bridges. Therefore, in order to improve the practicality of the physically-informed VINO, data-informed fine-tuning is carried out using available data from real bridges. It is noted that no damage data is required in the fine-tuning stage. The feasibility of the drive-by bridge damage detection using VINO is discussed using an in-house experiment on a model bridge under a moving vehicle.

Numerical benchmark for road bridge damage detection from passing vehicles responses applied to four data-driven methods

Drive-by bridge monitoring utilizes measured responses from passing vehicles to perform damage detection of bridge, a methodology challenged by multiple factors and operational conditions. Recently, data-driven methods have been used to improve the accuracy of drive-by monitoring. This thriving research field requires (but lacks) publicly available datasets to improve and validate its monitoring and damage detection capabilities. To foster data-driven drive-by bridge damage assessment methods, this document presents an openly available dataset consisting of numerically simulated vehicle responses crossing a range of bridge spans with various damage conditions. The dataset includes results for different monitoring scenarios, road profile conditions, vehicle models, vehicle mechanical properties and speeds. The intention is to provide a useful resource to the research community that serves as a reference set of results for testing and benchmarking new developments in the field. In addition, four recently published data-driven drive-by methods have been tested using the same dataset.

Drive-by scour damage detection in railway bridges using deep autoencoder and different sensor placement strategies

Drive-by scour damage detection in railway bridges using deep autoencoder and different sensor placement strategies

Identification of bridge influence line and multiple-vehicle loads based on physics-informed neural networks

Influence lines (ILs) and vehicle loads identification are critical in the design, health monitoring, and damage detection of bridges. Traditionally, the approach used in most existing literature has been to solve the system of equations directly. However, these approaches require complex calculations such as matrix decomposition and regularization coefficient optimization, making them difficult to implement. In addition, there are difficulties in obtaining accurate axle information and effectively separating the bridge response due to each vehicle. Thus, the improvement of identification algorithms for ILs and multi-vehicle loads remains of significant importance. To address these issues, this paper presents a novel approach that integrates prior physical equations and neural networks. This is achieved by integrating the equation that reflects the relationship between axle loads and bridge response into the neural network, utilizing existing methods for acquiring axle information of vehicles. To validate the effectiveness of the proposed method, it was first applied to theoretical and simulation data. The study then investigated the impact of noise and dynamic effects on the accuracy of the results, as well as the range of the neural network layers and sampling intervals. Finally, the method was implemented for identifying multiple-vehicle loads. The findings of the study confirm the feasibility and numerical stability of the proposed approach. The proposed method eliminates the need for complex computational processes, including matrix decomposition, diagonalization, regularization coefficient optimization, and solution vector smoothing fitting. As a result, the implementation of the algorithm is significantly less challenging, and identification accuracy is improved. It is important to note, however, that the proposed method is relatively more time-consuming due to the iterative learning and training required by the neural network.

Application of variational mode decomposition in vibration-based bridge damage detection

Bridges are critical infrastructure that require continuous structural health monitoring to ensure safety. Vibration-based methods use changes in dynamic response for damage detection. In this study, variational mode decomposition (VMD), a signal analysis technique, was applied to detect damage in a three-span beam bridge model subjected to moving loads. A finite-element model was developed to represent the bridge–vehicle interaction and damage was simulated by reducing the flexural stiffness of the bridge. Sensors placed on each span recorded the vibrations during vehicle passage. VMD was used to decompose the signals for analysis of damage indications. Different damage levels (30–50% stiffness reduction) and locations (each span) were evaluated. The vehicle speed was also varied to assess its effect. Damage was identified as peaks in the instantaneous energy diagrams of decomposed signals, correctly indicating the axle locations. Detection was improved with sensors nearer the damage and lower vehicle speed. The results demonstrate the ability of VMD to detect and locate damage without a baseline, indicating its potential as a non-baseline vibration-based damage detection method for bridges.

Bridge Damage Detection with Support Vector Machine in Accelerometer-Based Wireless Sensor Network

Abstract Purpose This paper proposes an in-network vibration data processing using Wireless Sensor Network (WSN) leveraging Machine Learning (ML) for damage detection and localization. The study also presents the ML algorithms comparison that is suitable to be deployed in WSN and implemented the proposed cluster-based WSN topology on the bridge simulation test. Methods The bridge vibration data was acquired using accelerometer-based wireless sensor nodes. The data collected are transformed using Fast Fourier Transform (FFT) to obtain fundamental frequencies and their corresponding amplitudes. The machine learning method i.e., Support Vector Machine (SVM) with linear and Radial Basis Function (RBF) kernel was used to analyze the vibration data collected from the WSN. In-network data processing and cluster-based WSN topology is implemented and the programmable wireless sensor nodes is utilized in this study. Results The experiments were conducted using real programmable wireless sensor nodes and developed our test bed bridge which makes this work different from the previous studies. The classification and predicting results shows 97%, 96%, 97%, and 96% for accuracy, precision, recall rate, and f1-score, respectively. Conclusion Machine learning methods can potentially be combined with the vibration WSN for bridge damage detection and localization.

Train-induced vibration response reconstruction for bridge damage detection with a deep learning methodology

Sensor malfunctions in bridge structural health monitoring system (BSHMS) could cause the missing of measured data, reducing the accuracy of damage detection. It is of great significance to reconstruct the data of faulty sensors with measured data of well-functioning sensors for bridge structural state assessment. In this paper, a train-induced vibration response reconstruction approach for bridge damage detection is proposed. An enhanced deep densely connected neural network (ED-DCNN) introducing a multi-head self-attention mechanism and a skip connection is designed to reconstruct the missing bridge responses, aiming to improve the vibration correlation between the faulty sensors and good sensors and benefit the learning of temporal and spatial features of responses. Then to further validate the accuracy and reliability of the ED-DCNN, an alternate down-sampling residual network (AD-ResNet) is built to identify bridge damage with the complete responses, including the reconstructed ones and the measured, which extracts the structural damage features at different scales utilizing different types of residual blocks in an alternating manner. The efficiency and robustness of the two networks are comprehensively studied using the collected acceleration data from a heavy-haul railway bridge, considering different noise levels. The results demonstrate that under various conditions, the reconstructed responses match the real responses well in time domain, with a low RMSE below 0.4% and a high correlation coefficient of 0.9744, which indicates that the incorporation of a multi-head self-attention module effectively improves the construction of correlations between responses. Additionally, the natural frequencies of bridge are accurately identified from the reconstructed responses. In terms of bridge damage detection, taking both the measured and reconstructed responses as inputs, the identification accuracy obviously improves by 33.57% compared to cases without the reconstructed responses, further emphasizing an excellent reconstruction capability of the ED-DCNN.

Ensemble learning-based structural health monitoring of a bridge using an interferometric radar system

Due to the increase in population, urbanisation, transportation development, and the aging of existing bridges, there is a growing need for new and rapid structural health monitoring (SHM) of bridges. To address this challenge, a method that stands out is the use of an interferometric radar system-based device, specifically Image by Interferometric Survey-Frequency for structures (IBIS-FS). Known for its portability and non-intrusive operation, IBIS-FS does not require direct contact with the bridge. This study utilised IBIS-FS to capture a pedestrian bridge’s natural frequencies and mode shapes. The data obtained were found to be consistent with results from finite element models, demonstrating the reliability of IBIS-FS in capturing modal parameters. Building upon this foundation, the study then explores the application of advanced ensemble-based machine-learning techniques. By leveraging the data acquired from IBIS-FS, algorithms such as Random Forest, Gradient-boosted Decision Trees (GBDT), and Extreme Gradient Boosting (XGBoost) are used for bridge damage detection. These machine-learning (ML) techniques are suited to analyse the incomplete modal parameters of bridges, as captured by IBIS-FS. The study focuses on using these algorithms to interpret the changes in modal parameters, specifically identifying damage as a reduction in the stiffness of elements. This approach allows for a comprehensive analysis, where the modal parameters, including mode shapes and natural frequencies altered by varying noise levels, are fed as input to the models. It was observed that all three ML methods, with Random Forest in particular, can effectively identify the location and severity of damage, demonstrating an efficient training process. The robustness of GBDT and XGBoost in handling complex data sets also shows great promise for their application in bridge damage detection. Collectively, these results underscore the potential of combining advanced ML techniques like Random Forest, GBDT, and XGBoost with the data acquired from IBIS-FS.

Structural Damage Detection Based on the Correlation of Variational Autoencoder Neural Networks Using Limited Sensors.

Identifying the structural state without baseline data is an important engineering problem in the field of structural health monitoring, which is crucial for assessing the safety condition of structures. In the context of limited accelerometers available, this paper proposes a correlation-based damage identification method using Variational Autoencoder neural networks. The approach involves initially constructing a Variational Autoencoder network model for bridge damage detection, optimizing parameters such as loss functions and learning rates for the model, and ultimately utilizing response data from limited sensors for model training analysis to determine the structural state. The contribution of this paper lies in the ability to identify structural damage without baseline data using response data from a small number of sensors, reducing sensor costs and enhancing practical applications in engineering. The effectiveness of the proposed method is demonstrated through numerical simulations and experimental structures. The results show that the method can identify the location of damage under different damage conditions, exhibits strong robustness in detecting multiple damages, and further enhances the accuracy of identifying bridge structures.

An unsupervised machine learning approach for real-time damage detection in bridges

The bridge network is progressively aging, with an alarming proportion of bridges over 100 years. This situation engenders substantial risks to the overall reliability of transportation networks, requiring innovative methods for efficient management. Monitoring can provide a direct source of information about structural behavior generating alerts when changes occur. Real-time alerts enable effective infrastructure management and decision-making during damage or anomalous situations. However, monitoring can result in a large amount of data that is often difficult to convert into valuable information in real time. This paper presents an approach for real-time detection of abrupt damage occurrence in bridges using unsupervised anomaly detection algorithms and strain/acceleration measurements. The approach incorporates the separation of measurements into events having the same loading nature and the construction of three feature matrices based on statistical features, time-frequency features, and wavelet spectrum features. It includes the evaluation of five anomaly detection algorithms including Isolation Forest, One-Class Support Vector Machine, Robust Random Cut Forest, Local Outlier Factor, and Mahalanobis Distance. The approach is illustrated with a case study of a steel-bascule-railway bridge, that has experienced a brittle cracking event during monitoring. Results highlight the robustness of One-Class Support Vector Machine, Isolation Forest, and Local Outlier Factor algorithms in promptly detecting abrupt changes across different features. The separation of strain and acceleration data into loading-based events, coupled with the comparison of previous and new event features, provides robust feature matrices for effective damage detection. Enhanced detection and higher scores are particularly attributed to time-frequency domain features during damage occurrence. The presented approach can be used as a base on how to perform real-time anomaly detection within the context of bridge monitoring.

Mode Shape-Based Damage Detection of Twin Skewed Bridges Based on OMA and FE Model Updating

A comparative study of the dynamic behavior of twin single-span bridges was carried out based on operational modal analysis and a linear finite element model updating with the aim of finding a potential damage indicator. The bridges are skewed, 30 m long and with prestressed concrete girders as the main support structure, but one of them were collided and suffered damage in two of its I-shape girders. Ambient vibration tests were conducted by deploying 14 uniaxial accelerometers along the sidewalks, experimental modal parameters were identified using the COVariance-driven Stochastic Subspace Identification method (SSI-COV). Six vertical modes were clearly identified for each structure. Additionally, some extra modes with non-proportional damping were identified. It was found that the modal frequencies are sensitive to the support conditions and the overall stiffness of the bridge but not to the localized girder damage, which, on the contrary, is reflected in the discrepancy in the fundamental bending mode shape amplitude between the lateral sides of the bridge, and a so-called Continuous Symmetry Measure adopted as a symmetry index seems to be a good damage indicator for this type of straight-alignment simply supported bridges, given the symmetric nature of the bending mode shapes and the asymmetric nature of the damage.

FE Model Updating of Continuous Beam Bridge Based on Response Surface Method

In this paper, A high-order response surface method is proposed for finite element model updating of continuous beam bridges. Firstly, based on visual inspection and environmental vibration testing, the peak picking (PP) method and random subspace identification (SSI) method are used to identify the dynamic characteristic parameters of the structure. Then, the finite element model of the continuous beam bridge is updated based on the third-order response surface method. It can be concluded that the results of the updated finite element model are in good agreement with the test results, and the maximum error between the calculated and measured frequency is less than 3%, with MAC values greater than 85%. Moreover, the updated finite element model can reflect the current situation of real bridges and serve as the basis for bridge health monitoring, damage detection, and safety assessment.

Drive-by Bridge Damage Detection Using Continuous Wavelet Transform

Bridges serve as vital engineering structures crafted to facilitate secure and effective transportation networks. Throughout their life-cycle, they withstand various factors, including diverse environmental conditions, natural hazards, and substantial loads. Recent bridge failures underscore the significant risks posed to the structural integrity of bridges. Damage detection techniques, being core components of structural health monitoring, play a crucial role in objectively assessing bridge conditions. This article introduces a novel framework for identifying damage in bridges utilizing continuous wavelet analysis of accelerations recorded using two sensors mounted on a vehicle traversing the bridge. The proposed method leverages changes in the static response of the bridge, which has proven to be more sensitive to damage than its dynamic counterpart. By doing so, the method eliminates the reliance on modal parameters for damage identification, addressing a significant challenge in the field. The proposed framework also addresses key challenges encountered by drive-by monitoring methods. It mitigates the adverse effects of road roughness by utilizing residual accelerations and efficiently detects and locates damage even in the absence of corresponding data from an undamaged bridge. Numerical investigations demonstrate the robustness of the proposed method against various parameters, including damage location and extent, vehicle speeds, road roughness levels, different boundary conditions, and multi-damage scenarios.