Bridge Health Assessment Research Articles

Anomaly detection for bridge health monitoring data based on multiple encoded images and convolutional neural network

In bridge health monitoring systems, abnormal monitoring data inevitably appear due to the complex operational conditions, and these anomalous data will seriously affect the accuracy and reliability of the results of bridge health assessment. Hence, it is vital to detect the abnormal data before further processing. However, the pattern diversity and sample scarcity usually make the accuracy of data anomaly detection unsatisfactory. A novel anomaly detection method for bridge health monitoring data based on encoded images and convolutional neural network (CNN) is proposed in this paper. First, the raw time series are encoded into images to highlight the intrinsic features in the anomaly data, and multiple encoding techniques are adopted to represent the data in different perspectives. Then, the CNN is utilized to establish the mapping between the encoded images and the anomaly patterns for achieving the data anomaly detection. At last, the proposed method is verified with the acceleration data measured from an actual large span cable-stayed bridge. It shows that the proposed method can significantly improve the performance of data anomaly detection while the results are hardly affected by the small sample sizes, and a detection accuracy of over 95% is achieved based on the multiple encoded images.

Thermography measurement for bridge displacement in the darkness using power-free target

Vision-based displacement measurement has become increasingly popular in bridge health assessment. However, this technology faces challenge that a complicated power-consuming supply system must be installed in the darkness, making it unable to meet the needs of low energy consumption and low cost. This paper proposes a bridge displacement measurement method using power-free target and thermography to address it. First, a thermography measurement system for bridge displacement in the darkness is developed. The power-free target with a hollowed-out circular feature is adopted to form the contrast to the background, and graphene is coated on the target surface to enhance the infrared emissivity and contrast further. Second, a sub-pixel elliptic center positioning algorithm based on thermography enhancement and data optimization is proposed to improve the accuracy and the ability to resist occlusion. The feasibility of the method is verified by laboratory tests and bridge field tests. When the measurement distance is within 15 m, the results meet the engineering error requirements of 5%, with the prospect of popularization in small and medium-sized bridges displacement measurement.

Damage Identification Method of Tied-Arch Bridges Based on the Equivalent Thrust-Influenced Line

Early tied-arch bridges cannot satisfy the current traffic load demand due to their load grades and maintenance levels. Also, these tied-arch bridges have accumulated structural damage with increasingly prominent safety risks. In order to accurately evaluate the characteristics of tied-arch bridges’ structural-influenced lines and identify damage of their arch ribs and suspenders, two analytical solutions are derived and established in this paper for the thrust-influenced lines of parabola and catenary two-hinged arches with a tie beam. A new method and an index for identifying damage of arch ribs and suspenders of tied-arch bridges are proposed. The results of numerical simulation in this paper verified that the proposed analytical solution has good analytical accuracy in practice on those two-hinged nonflat arches. With the use of equivalent thrust-influenced line difference curvature, the effectiveness of damage identification and the verification method were verified on the tie beam, suspender, and arch rib of plane tie arch structure as well as the suspender and arch rib of tied-arch bridges in this research. Furthermore, combining with grey relation analysis, the noise immunity of the proposed index method can be verified. Also, thrust-influenced line recognition based on VMD (variational mode decomposition) is introduced, and a practical process of bridge health assessment based on the quasistatic-influenced line loading of three-axle heavy vehicles is established. Theoretical analysis and numerical verification show that the calculated error of the analytic solution of two-hinged arches with a rise-span ratio greater than 1/8 is less than 9.57%, and the error decreases with the increase of the rise-span ratio. Therefore, it can be applied to the calculation and analysis of tied-arch bridges with a rise-span ratio range between 1/4 and 1/5. With the equivalent thrust-influenced line index proposed in this paper, the damage of suspenders and arch ribs of tied-arch bridges can be accurately located. It has been found that the proposed method is more effective to identify the damage of suspenders than the damage of arch ribs does, showing good noise immunity. In summary, this research has provided theoretical support for arch bridge design and evaluation. Combining with existing bridge-monitoring methods, the new bridge damage identification method proposed in this paper has the prospect of realizing the normal health status assessment of existing tied-arch bridges.

Integration, harmonization, and processing of geomatic data for bridge health assessment: the Lastra a Signa case study

A visual inspection, which entails field surveying, such as photodocumentation and footage, is the first step of a multi-level approach to bridge health assessment. Furthermore, the use of surface models, CAD drawings, and orthophotos, guarantees complete and accurate documentation, thus allowing for a better understanding of the environment, the anthropic structures, and their relationships. All the georeferenced surveys’ outputs are especially advisable within a prespective of periodical monitoring, as forseen by national legislation. The work is aimed to study two adjacent bridges over the Arno River in Lastra a Signa, Italy. One of the bridges, Ponte nuovo sull’Arno, is an overpass for motor vehicles. The other one, which is called Passarella sull’Arno, is a pedestrian viaduct. A topographic reference network has been settled using the GNSS survey technique. 3D point clouds of the bridges have been acquired by performing a Laser Scanning survey. A bathymetric survey has been carried out to acquire a 3D point cloud of submerged bridges’ parts and the riverbed. Through a Photogrammetric survey from RPAS, an orthophoto of the area has been built. Finally, evidence of historical submerged bridge structures has been identified thanks to the multi beam survey. The work’s objective is to integrate surveying geomatics techniques to create a reliable survey of the bridges, the surrounding area, and the riverbed, as support to the most common structural health assessment methods.

Autoencoders for unsupervised real‐time bridge health assessment

AbstractOver the last decades, the rising number of aging infrastructures has progressively fueled much interest toward the field of structural health monitoring. Following the increasing popularity of artificial intelligence algorithms, an autoencoder‐based damage detection technique within the context of unsupervised learning is proposed in this paper to provide support for practical engineering applications. The developed methodology uses the autoencoder to reconstruct raw acceleration sequences of user‐defined length collected from a healthy structure. To quantify the errors between the original input and the reconstructed output, which may be representative of damage occurrence, two indexes of reconstruction loss are selected as damage‐sensitive features. To support damage detection, a selected number of short‐time sequences are finally grouped into a unique macrosequence. The novel procedure can effectively both work at the single sensor level, as well as combine the predictive models using an ensemble learning strategy. Avoiding system identification, results obtained in the Z24 bridge demonstrate that the proposed method is quite effective for local damage detection with limited computational effort and using a limited number of sensors, thereby suitable to be easily applicable in the context of real‐time bridge assessment.

Effective multilayer hybrid classification approach for automatic bridge health assessment on large-scale uncertain data

The health level of the bridge is critical to the safety and maintainability of the bridge. However, with the rapid increase of bridge data with complex information, manual evaluation of bridge health requires high labor and time costs and a lot of related knowledge. Meanwhile, due to the error of sensors and subjective judgment of experts, as well as the influence of the external environment of the bridge, there is great uncertainty in the evaluation of the bridge data. Thence, how to use a large amount of bridge data with large-scale uncertain labels to build a robust classification model for efficient and high-quality bridge health assessment has become an urgent task. In order to better assess the health of the bridge, we adopt a multi-layer hybrid method to iteratively determine the uncertain labels of the target dataset, evaluate the confidence of the large-scale uncertain labels, add high-confidence data to the training set, and correct the low-confidence data. Finally, we get an effective classification model with the optimized training dataset. This paper studies the learning problem of classification model on labeled data with large-scale uncertain labels, and proposes an effective hybrid classification model (HCM), which can establish a supervised classifier under the condition of detecting uncertain labels and realize error label correction. In order to measure the HCM label assignment problem, we introduce a new penalty function, which can evaluate the label consistency problem of two basic classifiers. Meanwhile, we apply the model to bridge data with uncertain labels for bridge health evaluation. Experiments conducted on synthetic data, benchmark data and real bridge datasets show that the proposed method is superior to other methods and provides an effective and convenient solution for bridge health assessment. At the same time, this method can also be used in other research fields where there are large-scale uncertain labels.

Dynamic monitoring of bridges: accelerometer Vs microwave radar interferometry (IBIS-S)

Dynamic monitoring of bridge structures require an innovative technique which could potentially be used for rapid damage assessment and provide reliable information. This paper presents a comparison study on bridge dynamic monitoring using accelerometer and microwave radar interferometry (IBIS-S) for bridge health assessment. In particular, this study describes the advantages and disadvantages of the measurement used in the recent studies. The results of study will provide preliminary information for selecting of the appropriate methods related to dynamic monitoring of bridges.

Automatic detection method of bridge cracks based on residual network

At present, the level of bridge crack detection is not high. In order to solve the problems of weak robustness of traditional image processing methods and time-consuming and laborious manual detection, a pixel level detection algorithm of bridge cracks based on adding residual block is proposed, which provides more intuitive and accurate detection results for bridge health assessment. Firstly, the concrete beam slab crack data set with a certain amount of data is built to train and test the model. After comparing and analyzing the four learning rates, 2×10−4 is chosen as the initial learning rate. Finally, in the aspect of fracture segmentation, the model proposed in this paper has good performance and good crack identification effect. It can accurately and efficiently locate, classify and identify fractures, which has certain practical application value.

An End-to-End Framework With Multisource Monitoring Data for Bridge Health Anomaly Identification

Bridge health assessment has been a challenging problem due to great evaluation errors caused by heterogeneous and different dimensional bridge factors. Most approaches employ monitoring data of one bridge factor to simplify the evaluation problem resulting in poor assessment performance. To address this issue, this article proposes an end-to-end framework to evaluate the health of bridges by exploring objective features and correlations of multiple monitoring factors. This model aims to learn representative features from raw monitoring data of bridge factors (i.e., strain, temperature, traffic flow, and heavy vehicle number) and classify the comprehensive features into different health degrees in a single framework. Especially, in terms of characteristics of bridge factors (e.g., time sequence and heterogeneity), a hierarchical learning structure with multiple convolutional, pooling, and dense layers is presented to learn the representations of the whole bridge monitoring data. The structure contributes to capturing rich information from different observed factors and improving feature learning ability through designing particular neural networks for each bridge factor in accordance with their corresponding data structure. In addition, a classification scheme with multiple fully connection layers and support vector machines in the framework is designed to achieve higher evaluation accuracy. Experiments on real-world monitoring data of a specific bridge validate the superiority of the proposed model.

Concrete bridge damage detection using parallel simulation

Bridge health assessment is an interdisciplinary area of high socio-economical value. Bridge damage detection technology is one of the most developed technologies today. This paper introduces a new simulation-based damage detection method for reinforced concrete bridge structures using parallel processing. The proposed method takes advantage of vehicle axle loads and high accuracy bridge response from an advanced bridge monitoring system, executes simultaneously variant structural analyses with hypothetical damage assumption to reduce processing duration and improve productivity of damage detection, and utilises the mathematical model of Residual Sum of Squares in the comparison of simulation results and monitoring data to determine the best-fit models. In the case study, seven scenarios are simulated to explore and verify the features of the proposed method. A discussion of perspectives for further implementations completes the paper.

Wavelet Analysis With Different Frequency GPR Antennas for Bridge Health Assessment

Abstract This study presents an examination and a presentation of ground-penetrating radar (GPR) on a heavily used high-speed railway bridge in Turkey. This study also considers GPR mapping using two different antennas (2 GHz and 500 MHz), and their reliability in terms of locating subsurface features, such as rebar (lower and upper reinforcement) and the presence of moisture. Moreover, GPR interesting results in terms of the structural crack locations by detecting discontinuities and breakdowns in the wavelets' travel with the incorporation of wavelet analysis into the process of the collected raw data of the high-speed railway bridge under survey. This study promotes the GPR and the wavelet analyses in the health monitoring and assessment of bridge structures. By increasing insight into the kinds of structural deterioration that can occur in structures, such as high-speed railway bridges, this study also has implications for the process of tool selection in structural inspection. The wavelet analysis can be used as an effective tool in analyzing and presenting the results of the GPR survey.

Corroboration of NDT and deconvolution neural networks for pedestrian bridge health assessment

This paper describes the specific application of the non-destructive testing methods of visual inspection and ground penetrating radar (GPR) to a pedestrian bridge in Izmir, Turkey. The paper concentrates on the implementation of a deconvolution neural network (DNN) which is a procedure that employs neural network algorithms. By introducing collected GPR data to the DNN, the existence and location of cracks, rebar and moisture ingress on pedestrian pathways can reliably be located, thus providing superior information on which decisions relating to the functionality and life expectancy of a structure can be formulated. This study will be of benefit to engineers in providing a detailed and dependable assessment of the current state of structures such as pedestrian bridges.

Wireless sensor network protocol comparison for bridge health assessment

In this paper two protocols of Wireless Sensor Networks (WSN) are examined through both a simulation and a case study. The simulation was performed with the optimized network (OPNET) simulator while comparing the performance of the Ad-Hoc on demand Distance Vector (AODV) and the Dynamic Source Routing (DSR) protocols. This is compared and shown with real-world measurement of deflection from eight wireless sensor nodes. The wireless sensor response results were compared with accelerometer sensors for validation purposes. It was found that although the computer simulation suggests the AODV protocol is more accurate, in the case study no distinct difference was found. However, it was shown that AODV is still more beneficial in the field as it has a longer battery life enabling longer surveying times. This is a significant finding as a large factor in determining the use of wireless network sensors as a method of assessing structural response has been their short battery life. Thus if protocols which enhance battery life, such as the AODV protocol, are employed it may be possible in the future to couple wireless networks with solar power extending their monitoring periods.

Bridge Health Evaluation System based on the Optimal BP Neural Network

Neural network has strong ability of pattern recognition. In consideration of the problems of the traditional pure BP neural network, such as subjecting to the randomness of initial weights, slow convergence speed, low efficiency, easy to fall into local extreme value, in this paper we proposing an optimal BP network fusing with the genetic algorithm using in bridge health assessment. The optimized BP network algorithm has a good diagnosis effect, and improves the calculation accuracy and speed of the identification of bridge structure damage.

Analysis Of Bridge Safety Assessment With Correlation Between Measuring Points For Bridge Health Monitoring

Abstract By using linear statistical coefficients such as Pearson, Spearman and Kendall’s, and nonlinear methods such as time-delayed transfer entropy and mutual information, the correlation between the observation stations of bridge health monitoring system is obtained. A model for predicting structural safety based on the correlation coefficients is established, which provides a new research method for bridge health assessment and prediction based on bridge health monitoring system