Journal of Civil Structural Health Monitoring最新文献

Missing data due to sensor or transmission failures pose a significant challenge in structural health monitoring (SHM) systems, and data reconstruction methods can effectively address the missing data problem. Most of the traditional approaches typically focus on single-task data reconstruction, requiring repeated applications for each sensor and increasing computational cost. To address this issue, in this paper, we propose a probabilistic deep learning-based approach for multi-task data reconstruction. The multi-task data reconstruction is achieved by a probabilistic learning-based attentive neural process network (ANPN) that uses a common implicit data-driven kernel to learn the relationships among sensors. The meta-learning strategy is employed to train the common kernel in the ANPN. The attention mechanism is incorporated to further improve the reconstruction accuracy by enhancing the learning of the relationship between missing data and observed data. The effectiveness of the proposed ANPN is evaluated using the simulation data from a square pyramid space grid and the field data acquired from the Xiong’an Railway Station. The results show that the proposed ANPN can accurately reconstruct the missing data from multiple sensors within a second, underscoring its computational efficiency and accuracy. Furthermore, the influence of critical parameters (i.e., network depth, feature size, attention mechanism, and data loss ratio) on the reconstruction accuracy and efficiency is comprehensively investigated, and the optimal parameter settings are suggested.

Finite element method (FEM) is one of the essential means of structural analysis. However, the existing finite element modelling relies on manual and design drawings. Therefore, this study proposes an automated method for the numerical analysis of in-service bridges represented by point clouds. The proposed method includes two main innovations: first, an improved finite cell method (FCM) is introduced to generate finite element meshes from point clouds directly. This method eliminates the need for intricate computations involving uniformly distributed grid points as division criteria, significantly reducing the modelling time. Second, to overcome FCM’s limitations in handling structures with multiple material properties, this paper introduces a combination of a three-way topological relationship determination method (TRDM) and RandLA-Net. This approach automatically classifies material properties at integration points within the bridge structure’s physical domain. A model of an arch bridge is subjected to indoor experiments. Through comparative experimentation and ANSYS outcomes, proposed method demonstrates a level of precision akin to that of conventional modelling approaches.

Dams are a type of civil infrastructure that can directly impact people’s well-being, as their function is energy production, flood control, or water supply. Therefore, it is worth generating strategies to assess its current condition, since structural changes may occur during its useful life. One highly effective approach for evaluating the structural integrity of dams involves monitoring alterations in modal parameters. This method enables the identification of abnormal changes that may arise from structural degradation. Numerous studies have revealed the strong influence of environmental factors on modal parameters, resulting in variations unrelated to structural damage. This paper investigates the effects of environmental factors such as upstream water level and air temperature on the temporal evolution of the identified modal parameters of a hydroelectric dam’s hollow-gravity concrete block. Modal identification is performed through an automatic procedure of estimating modal parameters to 30-min acceleration time series over 3 years of operation. Correlation analysis reveals a distinct relationship between the identified modal parameters and environmental factors. Changes in air temperature exhibit a direct proportional impact on natural frequencies, while fluctuations of the upstream level have an inverse effect. Furthermore, a time lag was observed in the natural frequencies concerning air temperature. Multiple linear regressions were fitted to mitigate the induced effects, incorporating as predictors the upstream water level and the averages of air temperature segments measured prior to the predicted frequency. A reduction in variability of more than 50% was achieved in an out-of-sample 8-month period for the modes linked to the natural frequencies most influenced by environmental factors.

Accurate leak detection for water distribution networks (WDNs) is a critical task to minimize water loss and ensure efficient infrastructure management. Machine learning (ML) algorithms have demonstrated significant potential in establishing effective acoustic leak detection systems. However, the utilization of time-series models, specifically designed to handle sequential signals, in the field of water leak detection remains relatively unexplored, and there is a lack of research discussing their applicability in this context. Therefore, this study introduces a novel approach for precise leak detection in WDNs using a Time-Transformer model, which effectively captures long-range dependencies through self-attention mechanisms, enabling it to outperform other time-series models. This study conducted field experiments on WDNs in Hong Kong to demonstrate the superior performance of the proposed approach in accurately detecting leaks. The model structure is optimized through parametric experiments. Besides, leak detection and t-SNE results highlight the model's significant potential to enhance leak detection in WDNs compared to 1D-CNN and CNN–LSTM. The proposed Transformer-based model shows significant potential in advancing leak detection in WDNs, improving accuracy and precision, and supporting efficient water management.

Existing bridges constitute essential infrastructures of land transport and communications routes worldwide. They are often outdated and vulnerable; for this reason, monitoring and safety should be ensured for their use. The reduced economic and technical resources lead to the necessity of defining intelligent monitoring strategies for the preliminary classification of the infrastructures to establish an order of priority for executing more in-depth checks, verifications, and interventions. In this context, earth monitoring through satellite remote sensing has become a fundamental research topic in the last decades. This technique allows to obtain innumerable information on the temporal and spatial evolution of displacements at a territorial scale by means of the observation of wide deformation phenomena such as subsidence, landslides, and settlements. Furthermore, at a smaller scale, as in the case of a single bridge, the use of high spatial resolution and high sampling rate data could be crucial in civil engineering scenarios to carry on a preliminary structural monitoring of a road, railway network, or a single bridge. This work proposes a procedure for a large-scale analysis for the monitoring of an entire road network, based on remote sensing Structural Health Monitoring (SHM). The capability of the procedure is investigated on a network of 68 bridges, using deformation measurements derived from satellite remote sensing, where large stacks of ascending and descending Differential SAR Interferometry DInSAR data products were available. A Risk Class is estimated for each bridge based on the deformation analysis, considering the potential phenomena at both territorial and local scales. Based on such a Risk Class, the stakeholders can define most critical bridges as well as more in-depth monitoring strategies.

Bolted steel plate joints encounter challenges posed by joint corrosion, which impact the quality of interfacial contact among the bolted components. Unfortunately, no correlation between corrosion-induced joint damage and preload, nor an existing numerical model capable of capturing such effects, has been identified. This study aims to utilize guided wave ultrasonic investigation to examine the deterioration of interfacial contact caused by corrosion in bolted joints. Additionally, a contact modification-based numerical approach is presented to capture the effects of changing interfacial stress during joint corrosion. Guided wave mode selection was conducted with some preliminary experiments supplemented with the theory of wave mode dispersion, hence leading to the selection of S0 and A0 mode existing at 300 kHz. The joint was then corroded in a controlled manner using an electrochemical process while simultaneous ultrasonic measurements were taken. The experimental observations highlighted the progressive dispersion in the transmitted A0 mode across the bolted joint, potentially due to changing interfacial stress boundaries between the plates. A damage parameter, termed the dispersion index, was developed based on the energy ratio of different signal sections. A linear change in the dispersion index was observed with the increase in corrosion-induced mass loss. The insight was further established through a numerical investigation by studying the effect of changing bolt preload and the corresponding interfacial stress distribution. The findings revealed that monitoring the changes in the stress distribution at the bolted interface can provide insight into interfacial corrosion. Eventually, destructive tension test results confirmed the effect of joint corrosion on the load-bearing capacity of the joint. The change in failure mode of the pristine and corroded specimen is observed. The reported approach establishes the potential of ultrasonic inspection to investigate the interfacial health of a bolted joint in corroding conditions.

Bridges in service are subjected to environmental and load actions, but their status and conditions are typically unknown. Health monitoring systems have been installed on long-span bridges to monitor their loads and the associated responses in real time. Since 1997, the Tsing Ma suspension bridge in Hong Kong has been the world’s first of the type equipped with a long-term health monitoring system. For the first time, this study reports the first-hand field monitoring data of the bridge from 1997 to 2022. The 26-year data provide an invaluable and rare opportunity to examine the long-term characteristics of the loads, bridge responses, and their relationships, thereby enabling the assessment of the bridge’s load evolution and structural condition over time. Results show that traffic loads have remained stable after 2007, highway vehicles kept increasing until the COVID-19 pandemic in 2020, the annual maximum deck temperature continued to increase at a rate of 0.51 °C/decade, typhoon durations increased by 2.5 h/year, and monsoon speeds decreased and became dispersed and variable. For the bridge responses, deck displacement is governed by the varying temperature. Natural frequencies in the past 26 years were almost unchanged. The overall condition of the bridge is very satisfactory. Current status and recent update of the health monitoring system are also reported. Lastly, prospects of bridge health monitoring are discussed. This study is the first to report the over one-quarter century status of a structural health monitoring system and the behavior of a long-span suspension bridge. This research provides a benchmark for many other bridge monitoring systems worldwide.

Predicting bridge deflection is crucial for identifying potential structural issues, as sustained deviations from the expected range may indicate stiffness degradation. To address the stochastic modeling errors often overlooked by existing methods, this paper proposes a Bayesian Dynamic Noise Model (BDNM) for predicting the daily average deflection of bridge structures. The dynamic noise equations are formulated based on measured deflection data and incorporate modeling errors. Using Bayes’ theorem, a recursive BDNM process for bridge deflection prediction is established. Within a Bayesian forecasting framework, key parameters, particularly the coefficient and variance of modeling errors, are estimated using the method of moments, while the Bayesian discount factor is determined using Bayesian optimization. In addition, a novel prediction interval formula is developed, considering both modeling errors and monitoring uncertainties, based on the additivity of the normal distribution. This prediction interval is used as an anomaly detection threshold, and the estimated modeling errors from within the model are employed as damage indicators. The model is validated using monitoring data from an in-service bridge and compared with several common methods. Results demonstrate that the proposed method achieves high prediction accuracy and provides reasonable prediction intervals. Simulated scenarios of increased response variability due to stiffness degradation further illustrate the model’s sensitivity to structural behavior anomalies. This method lays a theoretical foundation for developing real-time warning systems for in-service bridges.

Inspections and condition monitoring of the stormwater pipe networks have become increasingly crucial due to their vast geographical span and complex structure. Unmanaged pipelines present significant risks, such as water leakage and flooding, posing threats to urban infrastructure. However, only a small percentage of pipelines undergo annual inspections. The current practice of CCTV inspections is labor-intensive, time-consuming, and lacks consistency in judgment. Therefore, this study aims to propose a cost-effective and efficient semi-automated approach that integrates computer vision technology with Deep Learning (DL) algorithms. A DL model is developed using YOLOv8 with instance segmentation to identify six types of defects as described in Water Services Association (WSA) Code of Australia. CCTV footage from Banyule City Council was incorporated into the model, achieving a mean average precision (mAP@0.5) of 0.92 for bounding boxes and 0.90 for masks. A cost–benefit analysis is conducted to assess the economic viability of the proposed approach. Despite the high initial development costs, it was observed that the ongoing annual costs decreased by 50%. This model allowed for faster, more accurate, and consistent results, enabling the inspection of additional pipelines each year. This model serves as a tool for every local council to conduct condition monitoring assessments for stormwater pipeline work in Australia, ultimately enhancing resilient and safe infrastructure asset management.

Delamination in reinforced concrete (RC) bridge decks can degrade the serviceability of entire bridges, leading to concrete spalling and steel rebar corrosion and eventually becoming a safety concern. Drone-based infrared thermography (IRT) offers a promising tool for rapid assessment of bridge deck delamination compared to labor-intensive coring and visual inspection methods. However, the performance of passive IRT in detecting the delamination of RC bridge decks at its minimum depth and size (i.e., spall 25 mm or less deep or 150 mm or less in diameter) stipulated under a ‘fair’ condition state in the 2019 AASHTO Manual for Bridge Element Inspection has not been verified adequately. In this study, four RC slabs of identical design were cast with embedded thin foam sheets to simulate a wide range of delamination in thickness, size, spacing, and depth. Together, the four slabs form a representative RC deck of a mark-up bridge. Controllable indoor active IRT tests of individual slabs were conducted to detect and quantify the foams that serve as a ground truth for the performance of drone-based passive IRT for deck delamination detection on the mark-up bridge as the embedded foams may be displaced during concrete slab casting and the slab support is altered during erection. Statistical analysis was carried out on the thermal contrasts of both passive and active IRT tests on the four slabs to investigate the effects of delamination geometry and embedment depth. Both the active and passive IRT methods proved successful in localizing delamination and identifying its equivalent thicknesses of as low as 1.63 mm and a size (150 mm in length or 25 mm in depth) corresponding to the ‘fair’ condition state in the AASHTO Manual for Bridge Element Inspection.