Structural Health Monitoring Data Research Articles

On the fragility of non‐structural elements in loss and recovery: Field observations from Japan

AbstractThe role of non‐structural elements (NSEs) in the seismic performance of buildings has been highlighted in past years. Research studies following state‐of‐the‐art methodologies generally find that when the structural collapse is not of significant concern, NSEs tend to dominate the repair costs and financial investment required in a building. This paper examines field observations from interviews and data collected from commercial buildings via structural health monitoring (SHM) following the 2018 Osaka earthquake in Japan. It suggests that fragility functions used in current methodologies for estimating NSE damage may not be entirely representative of the in‐situ reality and possibly underestimate actual damage. Additionally, interviews with building owners/managers indicate that the alarm and financial impact of NSEs was not as critical as anticipated, with much of the observed NSE damage not of serious concern and tolerable in many cases. This article provides discussion and insight into possible causes for these differences before discussing how current methodologies can benefit from these observations. It is believed that many of the fragility functions currently used to estimate NSE damage may not be representative because of differences in installation conditions and loading protocols used in experimental testing, possible interaction with other elements, variability in the quality of workmanship during installation and possible wear, tear and degradation during service. On the other hand, it is seen how SHM data recorded during seismic events may provide valuable data for an alternative means to develop fragility functions. Furthermore, it is seen that when building recovery states (RSs) other than the implicitly assumed ‘full recovery’ state used in guidelines like FEMA P‐58 are explored, the role of NSEs in direct monetary losses significantly reduces. This coincides with the field observations in Japan regarding the impact of NSEs and supports the recent developments in functional recovery on what building owners and occupants are prepared to tolerate post‐earthquake. It indicates that when discussing the relative importance of different building performance groups, it is vital that the expected RS is also stated, as for most decision‐makers following major events, functionality rather than full recovery remains the primary goal; hence, repairs and proactive measures should bear this in mind for more effective use of resources.

A multi-layer ML model evolutionary paradigm for high-accuracy individual aircraft SHM

Evolutionary Machine Learning (EML) has emerged as a novel direction for providing robust methodologies for effectively tackling complex problems. However, most current research focuses on multi-objective optimization problems in machine learning (ML). There is limited research on the EML mechanism along the temporal application process, where a fixed machine learning model trained with the prior dataset will inevitably be not representative of the target problem with time. Taking the Structural Health Monitoring (SHM)-ML damage diagnosis model as an example, this paper proposes a novel multi-layer ML model evolutionary paradigm with multi-source data for high-accuracy individual aircraft SHM. The proposed paradigm considers three interconnected layers for the SHM-ML damage diagnosis model, including the dataset to establish the diagnosis model, the model structure and its parameter training, and mapping relationship used in the model to give the damage size. Besides, for a specific model of aircraft, there exist multiple stages from design to service, where multi-source of data is fused for the multi-layer evolution, enabling continuous improvement of damage diagnosis accuracy. Experimental validation is conducted on critical connection structures of aircraft, specifically the attachment lug structure, using the active guided wave SHM method. The results demonstrate the effectiveness of the proposed multi-layer ML model evolutionary paradigm.

Calibration strategies for long-term strain forecasting in high-rise building columns

Concrete exhibits long-term time-dependent behavior due to creep and shrinkage that impacts the safety and serviceability of high-rise buildings. These rheological effects are difficult to predict due to their random nature, dependence on environmental conditions, and loading history. Current methods include the use of creep and shrinkage tests of structure-specific concrete samples to update compliance and shrinkage, and sophisticated numerical models for prediction of the long-term structural behavior. However, creep and shrinkage tests are time-consuming, and simulation requires reliable numerical models and often proprietary solvers that are not available to the structural health monitoring (SHM) practitioner. Further, uncertainty propagation in complex numerical models is rarely seen in the relevant literature. In contrast, data-driven prediction methods using SHM data and simplified analytical models have shown to be successful for prestressed concrete bridges. In this work, we investigate calibration strategies of creep and shrinkage models using SHM data toward data-driven forecasting of long-term time-dependent behavior of high-rise buildings. A calibration strategy is identified that enables significant and consistent improvement of forecasting of long-term time-dependent behavior. It is also shown that continuous calibration can provide good predictions at least 30 days ahead. First-order analytical uncertainty propagation formulas are also provided. The calibration strategies are evaluated on data from two residential high-rise buildings in Singapore. Recommendations to the SHM practitioners are also given.

Early warning for abnormal strains of continuous bridges using a proposed temperature–strain mapping model

Bridge strain is an essential monitoring index for small to medium span bridges because abnormal strains could be attributed to bridge deterioration or damage. Thus, an early warning method for detecting abnormal strains is highly desired. In this study, a temperature–strain mapping model for predicting the temperature induced strains was proposed by fusing structural health monitoring data and deep learning neural networks. First, the sensitivity of temperature induced strains to structural damage was analyzed, and the thermal strain was chosen as a suitable indicator for bridge performance detection. Next, a high-precision prediction model was established to overcome the time-lag effect between the original temperature and temperature induced strain sequences. By comparing the model’s prediction with the identification of measurements, an early warning method for detecting the abnormal temperature induced strains was proposed. The evaluation results demonstrated that the proposed early warning method was able to accurately detect the abnormal strains, even when the abnormal strain is smaller than the normal daily variations.

Similarity learning hidden semi-Markov model for adaptive prognostics of composite structures

Data-driven methodologies have found increasing usage in the last decade for remaining useful life (RUL) prognostics of composite materials utilizing structural health monitoring (SHM) data. Of particular interest is the reliable RUL prediction in cases where the end-of-life is not in between the extreme values within the testing dataset. For example, when unexpected phenomena that severely compromise the structural integrity occur during the service life. Such cases are often referred as outliers and the RUL prognosis based on a data-driven model that learns from past data is often erroneous. This study addresses this challenge by proposing a new stochastic model; the Similarity Learning Hidden Semi Markov Model (SLHSMM), an extension of the Non-Homogenous Hidden Semi Markov Model (NHHSMM). Through the utilization of a nonparametric discrete distribution, which characterizes the similarity between the testing structure and the training structures, a dynamic re-estimation process is employed. This process assigns higher importance to the training structures that display greater similarity to the testing one. As a result, the estimated parameters effectively capture the specific characteristics of the testing structure. The training and testing SHM data sets consist of strain measurements collected from a case study where carbon–epoxy single-stringered panels, are subjected to constant, variable, and random amplitude fatigue loading until failure. RUL estimations from the SLHSMM, the NHHSMM, and the Gaussian Process Regression (GPR) are compared. The SLHSMM clearly outperforms its classical counterpart and GPR providing more accurate outlier and inlier prognostics, demonstrating its capability to adapt to unexpected phenomena and integrate unforeseen data into a prognostic platform.

A systematic review of data fusion techniques for optimized structural health monitoring

Advancements in structural health monitoring (SHM) techniques have spiked in the past few decades due to the rapid evolution of novel sensing and data transfer technologies. This development has facilitated the simultaneous recording of a wide range of data, which could contain abundant damage-related features. Concurrently, the age of omnipresent data started with massive amounts of SHM data collected from large-size heterogeneous sensor networks. The abundance of information from diverse sources needs to be aggregated to enable robust decision-making strategies. Data fusion is the process of integrating various data from heterogeneous sources to produce more useful, accurate, and reliable information about system behavior. This paper reviews recent developments in data fusion techniques applied to SHM systems. The theoretical concepts, applications, benefits, and limitations of current methods and challenges in SHM are presented, and future trends in data fusion methods are discussed. Furthermore, a set of criteria is proposed to evaluate contents and information from original and review papers in this field, and a road map is provided discussing possible future work.

Directional Multi-Frequency Guided Waves Communications Using Discrete Frequency-Steerable Acoustic Transducers.

A novel directional transducer based on Guided Waves (GW) is introduced in this paper, designed for use in structural health monitoring (SHM) and acoustic data communication applications, i.e., systems in which the elastic medium serves as a transmission channel and information is conveyed through the medium via elastic waves. Such systems can overcome difficulties associated with traditional communication methods like wire-based or radio frequency (RF), which can be complex and have limitations in harsh environments or hard-to-reach places. However, the development of these techniques is hampered by GW dispersive and multi-modal propagation and by multi-path interference. The shortcomings can be effectively addressed by employing Frequency Steerable Acoustic Transducers (FSATs), which leverage their inherent directional capabilities. This can be achieved through the exploitation of a frequency-dependent spatial filtering effect, yielding to a direct correlation between the frequency content of the transmitted or received signals and the direction of propagation. The proposed transducer is designed to actuate or sense the A0 Lamb wave propagating in three orientations using varying frequencies, and has three channels with distinct frequencies for each direction, ranging from 50 kHz to 450 kHz. The transducer performance was verified through Finite Element (FE) simulations, accompanied by experimental testing using a Scanning Laser Doppler Vibrometer (SLDV). The unique frequency-steering capability of FSATs is combined with the On-Off Keying (OOK) modulation scheme to achieve frequency directivity in hardware, similar to ongoing research in 5G communications. The MIMO capabilities of the transducer were finally tested over a thin aluminum plate, showing excellent agreement with the FE simulation results.

Data Anomaly Detection for Structural Health Monitoring Based on a Convolutional Neural Network.

Structural health monitoring (SHM) has been extensively utilized in civil infrastructures for several decades. The status of civil constructions is monitored in real time using a wide variety of sensors; however, determining the true state of a structure can be difficult due to the presence of abnormalities in the acquired data. Extreme weather, faulty sensors, and structural damage are common causes of these abnormalities. For civil structure monitoring to be successful, abnormalities must be detected quickly. In addition, one form of abnormality generally predominates the SHM data, which might be a problem for civil infrastructure data. The current state of anomaly detection is severely hampered by this imbalance. Even cutting-edge damage diagnostic methods are useless without proper data-cleansing processes. In order to solve this problem, this study suggests a hyper-parameter-tuned convolutional neural network (CNN) for multiclass unbalanced anomaly detection. A multiclass time series of anomaly data from a real-world cable-stayed bridge is used to test the 1D CNN model, and the dataset is balanced by supplementing the data as necessary. An overall accuracy of 97.6% was achieved by balancing the database using data augmentation to enlarge the dataset, as shown in the research.

Vector Quantized Variational Autoencoder-Based Compressive Sampling Method for Time Series in Structural Health Monitoring

The theory of compressive sampling (CS) has revolutionized data compression technology by capitalizing on the inherent sparsity of a signal to enable signal recovery from significantly far fewer samples than what is required by the Nyquist–Shannon sampling theorem. Recent advancement in deep generative models, which can represent high-dimension data in a low-dimension latent space efficiently when trained with big data, has been used to further reduce the sample size for image data compressive sampling. However, compressive sampling for 1D time series data has not significantly benefited from this technological progress. In this study, we investigate the application of different architectures of deep neural networks suitable for time series data compression and propose an efficient method to solve the compressive sampling problem on one-dimensional (1D) structural health monitoring (SHM) data, based on block CS and the vector quantized–variational autoencoder model with a naïve multitask paradigm (VQ-VAE-M). The proposed method utilizes VQ-VAE-M to learn the data characteristics of the signal, replaces the “hard constraint” of sparsity to realize the compressive sampling signal reconstruction and thereby does not need to select the appropriate sparse basis for the signal. A comparative analysis against various CS methods and other deep neural network models was performed in both synthetic data and real-world data from two real bridges in China. The results have demonstrated the superiority of the proposed method, with achieving the smallest reconstruction error of 0.038, 0.034 and 0.021, and the highest reconstruction accuracy of 0.882, 0.892 and 0.936 for compression ratios of 4.0, 2.66, and 2.0, respectively.

Bridge Performance Prediction Based on a Novel SHM-Data Assimilation Approach considering Cyclicity

Modern bridges are monitored by an increasing network of sensors that produce massive data for bridge performance prediction. Reasonably and dynamically predicting with monitored data for the time-variant reliability of the existing bridges has become one of the urgent problems in structural health monitoring (SHM). This study, taking the dynamic measure of structural stress over time as a time series, proposes a data assimilation approach to predicting reliability based on extreme stress data with cyclicity. To this aim, the objectives of this article are to present the following: (a) a Gaussian mixture model-based Bayesian cyclical dynamic linear model (GMM-BCDLM) based on extreme stress data with cyclicity and (b) a dynamic reliability prediction method in the combination of GMM-BCDLM and SHM data via first-order second-moment (FOSM) method. An in-service bridge for providing real-time monitored stress data is applied to illustrate the application and feasibility of the proposed method. Then, the effectiveness and prediction precision of the proposed models are proved to be superior compared to other prediction approaches to extreme stress data with cyclicity.

Research on mapping relationship between environmental load and vibration response of bridge structure based on structural health monitoring data

AbstractWith the increasing lifespan of bridges, various forms of deterioration will inevitably occur. The bridge health monitoring system enables direct monitoring of the loads (such as wind, temperature, and vehicles) experienced by the bridge during operation, as well as the corresponding vibration response of the bridge structure (including main beam deformation, tower top displacement, cable acceleration, etc.). These data provide valuable insights into the safety and durability of the bridge during its operational lifetime. However, the challenge lies in extracting meaningful information from the vast amount of collected data. This article focuses on the Fuyu Bridge as a research case and analyzes the variations in environmental factors, such as temperature, humidity, wind load, and structural vibration responses (main beam and cable), obtained from the Fuyu Bridge's structural health monitoring system. It also conducts a structural safety assessment during the operational period of the bridge. Furthermore, the research explores the relationship between environmental factors, load effects, and structural response to establish a foundation for future structural safety assessments of the Fuyu Bridge during its operational period. For instance, changes in environmental factors or load effects can be used to predict structural responses. Additionally, this study provides access to the structural health monitoring data of the Fuyu Bridge, facilitating data‐driven research for other scholars in the field.

Stochastic stiffness identification and response estimation of Timoshenko beams via physics-informed Gaussian processes

Machine learning models trained with structural health monitoring data have become a powerful tool for system identification. This paper presents a physics-informed Gaussian process (GP) model for Timoshenko beam elements. The model is constructed as a multi-output GP with covariance and cross-covariance kernels analytically derived based on the differential equations for deflections, rotations, strains, bending moments, shear forces and applied loads. Stiffness identification is performed in a Bayesian format by maximising a posterior model through a Markov chain Monte Carlo method, yielding a stochastic model for the structural parameters. The optimised GP model is further employed for probabilistic predictions of unobserved responses. Additionally, an entropy-based method for physics-informed sensor placement optimisation is presented, exploiting heterogeneous sensor position information and structural boundary conditions built into the GP model. Results demonstrate that the proposed approach is effective at identifying structural parameters and is capable of fusing data from heterogeneous and multi-fidelity sensors. Probabilistic predictions of structural responses and internal forces are in closer agreement with measured data. We validate our model with an experimental setup and discuss the quality and uncertainty of the obtained results. The proposed approach has potential applications in the field of structural health monitoring (SHM) for both mechanical and structural systems.

Deformation warning index for reinforced concrete dam based on structural health monitoring data and numerical simulation

The material mechanical parameters of the dam body and foundation will change when a dam is reinforced during the aging process. This causes significant changes in the structural state of the project and makes it difficult to ensure its structural safety. In this study, a new deformation warning index for reinforced concrete dams was developed according to the prototype monitoring data, statistical models, three-dimensional finite element model (FEM) numerical simulation, and the critical conditions of the dam structure. A statistical model was established to separate the water pressure component. Then, a three-dimensional FEM of the reinforced concrete dam was constructed to simulate the water pressure component. Furthermore, the deformation components that affected the mechanical parameters of the dam under the same amount of reservoir water level change were separated and quantified accurately. In addition, the method for inversion of comprehensive mechanical parameters after dam reinforcement was used. The influence mechanisms of the deformation behavior of concrete dams under the reservoir water level and temperature changes were investigated. A new deformation warning index was developed by combining the forward-simulated critical water pressure component and temperature component in the period of extreme temperature decrease with the aging component separated by the statistical model. The new deformation warning index considers the structural state of the dam before and after reinforcement and links the structural strength criterion and the deformation evolution mechanisms. It provides a theoretical foundation and decision support for long-term service and operation management of reinforced dams.

A LCCA framework for the management of bridges based on data fusion from visual inspections and SHM systems.

AbstractBridges' integrity, safety and serviceability are fundamental targets that need to be ensured during the structural lifetime, owing to the crucial role that such infrastructure components play in modern communities. Recently, a particular attention is being paid to the management of bridges since many of them have reached the end of their expected service life. To this purpose, the Italian Ministry of Sustainable Mobility issued in 2020 the “Guidelines for risk classification and management, safety assessment and monitoring of bridges” that regulate visual inspections, risk assessment based on multilevel protocols and consequent monitoring and maintenance activities. In this context, structural health monitoring (SHM) systems play a fundamental role to guarantee safety and adequate performance levels. However, the synergy between visual inspections and SHM is yet to be fully understood and exploited. The main goal of this research work is to contribute to filling this gap by proposing a general probabilistic framework, based on Life‐Cycle Cost Analysis (LCCA), able to fuse information from SHM data and visual inspections, with the aim to plan over time short‐term and long‐term interventions and to optimize fund allocation by considering all the stored information.

Frequency-domain synchronization of structural health monitoring data

Since the advent of structural health monitoring (SHM), vibration analysis has been one of the most popular strategies for yielding modal information of structures, such as eigenfrequencies and mode shapes. As a result of the relatively high sampling rates adopted in vibration analysis, obtaining accurate modal information requires absolute synchronization of structural response data. Despite the advances in sensing technologies and synchronization protocols, synchronization discrepancies in structural response data may occur due to drifting in internal clocks of independent data acquisition units. Particularly in wireless SHM, in which each wireless sensor node essentially operates as an independent data acquisition unit, mitigating clock drifting has been an area of intensive research. This paper addresses the issue of data synchronization from a physics-based perspective, aiming to add an additional synchronization check at the post-sampling stage. Specifically, a frequency-domain synchronization approach is reported, which builds upon the relationship between time lags in structural response data and the cross spectral density phase angles. Validation tests are conducted using simulations of a multi-degree-of-freedom oscillator and data collected from a full-scale road bridge, demonstrating the capability of the approach to estimate time lags between SHM data even in the presence of non-proportional damping, thus overcoming a common limitation of other synchronization methods. The frequency-domain synchronization approach is intended to complement synchronization protocols in extracting accurate modal information.

A combined method of autoregressive model and matrix factorization for recovery and forecasting of cyclic structural health monitoring data

Structures such as bridges, offshore platforms, wind power facilities, and drilling rigs generate a corresponding periodic structural response under cyclic loads such as solar radiation, tides and sea-land breeze. Data recovery and forecasting are major parts of the work to ensure structural safety using periodic structural health monitoring data. In this context, a combined method of autoregressive (AR) model and matrix factorization (MF) method is presented to reaching the tasks of data imputation and prediction. The combined method with a graph-based temporal regularizer can well capture the cyclic characteristics and random character. The verification of the method is conducted using an in-situ temperature monitoring dataset of a high-speed railway steel bridge. The application of the method for both data recovery and forecasting with structured missing entries is carried out, which indicates the combined method owns outstanding performance in accurately recover missing entries and making predictions. In the last part, the analysis of rank order and length of time lag is made to show its influence on imputation accuracy.

Visualization of structural health monitoring information using Internet-of-Things integrated with building information modeling

Structural Health Monitoring (SHM) has become a paramount necessity in civil engineering for improving the operational performance of aging infrastructure. Recent monitoring techniques have utilized emerging technologies in Industry 4.0, such as the Internet of Things, Big Data analytics, cloud computing, and cybersecurity, to automate SHM methodologies. However, they have found challenges in linking these technologies and developing an autonomous, well-established digital framework for applications of SHM. Visualizing processed SHM data in a real-time digital interface generates multiple obstacles, such as witnessing delays in data transfer and resorting to offline tools for manual data processing. This paper, therefore, explores the integration of Building Information Modeling (BIM) and the Internet of Things (IoT) through an Arduino micro-processing unit for tracking and visualizing the data from the time and frequency domains. Strategies for enabling data monitoring and processing are developed while continuously acquiring structural responses. The query of data is established in a web-based database instead of storing the data in offline resources that await manual intervention. The proposed real-time SHM methodology is validated experimentally using two practical applications: a dynamically moving vehicle over a simply-supported bridge prototype and a randomly excited three-story model with real-time visualization of both time- and frequency-domain information under undamaged and damaged conditions. The proposed research develops an early-phase Digital Twin (DT) to present static and real-time dynamic data in a rich-fed BIM database.

Structural Condition Assessment and Temperature Effect Analysis of Cable–Pylon Anchorage Zone Using Long-Term SHM Data

To study the temperature effect in the anchorage zone of a long-span cable-stayed bridge, a wavelet multiresolution analysis and layered stripping method were used in this paper. Based on long-term structural health monitoring data, the signals were decomposed and reconstructed at multiple time scales, confirming temperature to be the main factor causing the stress change in the anchorage zone. The results of structural condition during operation showed that the general compressive stress level of the anchorage zone was low. However, the tensile stress level of the sidewall was high, which led to a severe concrete cracking. Excluding the influence of the seasonal temperature, the compressive stress increased slightly, the horizontal tensile stress in the upstream inner pylon wall increased, and the crack width increased gradually. By analysing the daily temperature effect on the upstream and downstream pylon walls, the regression model proposed in this study can be used to predict the daily temperature effect at any time in the diurnal cycle. The accuracy of the model is reliable within 6 days, but for the location of severe cracks, the monitoring data should be updated in real time to ensure the precision.

Vortex-Induced Vibration Recognition for Long-Span Bridges Based on Transfer Component Analysis

Bridge vortex-induced vibration (VIV) refers to the vertical resonance phenomenon that occurs in a bridge when pulsating wind passes over it and causes vortices to detach. In recent years, VIV events have been observed in numerous long-span bridges, leading to fatigue damage to the bridge structure and posing risks to driving safety. The advancement of technologies such as structural health monitoring (SHM), machine learning, and big data has opened up new research avenues for the intelligent identification of VIV in bridges. Machine learning algorithms can accurately identify the VIV events from historical data accumulated by SHM systems, thus providing an effective method for VIV recognition. Nevertheless, the existing identification methods have limitations, particularly in their applicability to bridges lacking historical VIV data. This study introduces an adaptive VIV recognition method in the main girders of long-span suspension bridges based on Transfer Component Analysis (TCA). The method can accurately identify VIV patterns in real-time or in historical data, even when specific VIV data are not available for the target bridge. The proposed method exhibits suitability for multiple long-span bridges. Experimental validation is performed using the SHM datasets from two long-span suspension bridges. The results show that the proposed VIV identification method can recognize more VIV samples compared to the benchmark model. When using sensor 1 data of bridge B as the source domain to identify the VIV of the L-section of bridge A, the F1 score of the TCA-based method is 0.836, while the F1 score of the benchmark model is 0.165. In the other 11 cases, the F1 score of the proposed model is higher than 0.8, which demonstrates the method’s robust generalization capabilities.

Multimodal Deep Neural Network-Based Sensor Data Anomaly Diagnosis Method for Structural Health Monitoring

Due to sensor failure, noise interference and other factors, the data collected in the structural health monitoring (SHM) system will show a variety of abnormal patterns, which will bring great uncertainty to the structural safety assessment. This paper proposes an automatic data anomaly diagnosis method for SHM based on a multimodal deep neural network. In order to improve the detection accuracy, both two-dimensional and one-dimensional features of the sensor data are fused in the multimodal deep neural network. The network consists of two convolutional neural network (CNN) channels, one a 2D-CNN channel for extracting time–frequency features of sensor data and the other a 1D-CNN channel for extracting raw one-dimensional features of sensor data. After convolution and pooling operations for the sensor data by the 2D channel and 1D channel separately, the two types of extracted features are flattened into one-dimensional vectors and concatenated at the concatenation layer. The concatenated vector is then fed into fully connected layers for final SHM data anomaly classification. In order to evaluate the reliability of the proposed method, the monitored data lasting for one month of a long-span cable-stayed bridge were used for training, validation, and testing. Six types of training conditions (missing, minor, outlier, over-range oscillation, trend, and drift) are studied and analyzed to address the issue of imbalanced training data. With an accuracy rate of 95.10%, the optimal model demonstrates the effectiveness and capability of the proposed method. The proposed method shows a promising future as a reliable AI-assisted digital tool for safety assessment in structural health monitoring systems.