Early damage detection under massive data via innovative hybrid methods: application to a large-scale cable-stayed bridge

Early damage detection is an initial step of structural health monitoring. Thanks to recent advances in sensing technology, the application of data-driven methods based on the concept of machine learning has significantly increased among civil engineers and researchers. On this basis, this article proposes a novel non-parametric anomaly detection method in an unsupervised learning manner via the theory of empirical machine learning. The main objective of this method is to define a new damage index by using some empirical measure and the concept of minimum distance value. For this reason, an empirical local density is initially computed for each feature and then multiplied by the minimum distance of that feature to derive a new damage index for decision-making. The minimum distance is obtained by calculating the distances between each feature and training samples and finding the minimum quantity. The major contributions of this research contain developing a novel non-parametric algorithm for decision-making under high-dimensional and low-dimensional features and proposing a new damage index. To detect early damage, a threshold boundary is computed by using the extreme value theory, generalized Pareto distribution, and peak-over-threshold approach. Dynamic and statistical features of two full-scale bridges are used to verify the effectiveness and reliability of the proposed non-parametric anomaly detection. In order to further demonstrate its accuracy and proper performance, it is compared with some classical and recently published anomaly detection techniques. Results show that the proposed non-parametric method can effectively discriminate a damaged state from its undamaged condition with high damage detectability and inconsiderable false positive and false negative errors. This method also outperforms the anomaly detection techniques considered in the comparative studies.

Wind-induced operational variability is one of the major challenges for structural health monitoring of slender engineering structures like aircraft wings or wind turbine blades. Damage sensitive features often show an even bigger sensitivity to operational variability. In this study a composite cantilever was subjected to multiple mass configurations, velocities and angles of attack in a controlled wind tunnel environment. A small-scale impact damage was introduced to the specimen and the structural response measurements were repeated. The proposed damage detection methodology is based on automated operational modal analysis. A novel baseline preparation procedure is described that reduces the amount of user interaction to the provision of a single consistency threshold. The procedure starts with an indeterminate number of operational modal analysis identifications from a large number of datasets and returns a complete baseline matrix of natural frequencies and damping ratios that is suitable for subsequent anomaly detection. Mahalanobis distance-based anomaly detection is then applied to successfully detect the damage under varying severities of operational variability and with various degrees of knowledge about the present operational conditions. The damage detection capabilities of the proposed methodology were found to be excellent under varying velocities and angles of attack. Damage detection was less successful under joint mass and wind variability but could be significantly improved through the provision of the currently encountered operational conditions.

In structural health monitoring, effectively eliminating the influence of variable environmental conditions on modal frequencies remains a critical challenge for accurate damage identification. Nonstationary and nonlinear variations in modal frequencies, commonly induced by environmental changes, tend to overshadow the effects caused by structural damage. An improved Gaussian mixture model (GMM) is proposed in this paper to normalize nonlinear and nonstationary frequency data, enabling effective structural damage detection under variable environmental conditions. As the effectiveness of the GMM is highly influenced by the initial parameter values used in the expectation-maximization (EM) algorithm, a subdomain division strategy is first presented to determine the unique initial values of the GMM parameters. Through the application of the EM algorithm, the GMM is constructed simply and efficiently through the determined initial parameters. Next, on the basis of the constructed GMM, the modal frequency data are normalized to extract damage features that remain unaffected by environmental variations. Subsequently, Hotelling’s T2 statistic and its cumulative form are calculated for the damage features and designated as the damage indicators; meanwhile, the corresponding damage thresholds are also calculated according to the kernel density estimation technique. To validate the proposed method, two case studies are conducted: one with a numerical mass-spring system and the other with a real bridge structure. Results show that environmental influences no longer impact the normalized frequency data, and the cumulative statistic demonstrates outstanding accuracy in identifying structural damage.

A wavelet-based approach is proposed for structural damage detection and health monitoring. Characteristics of representative vibration signals under the wavelet transformation are examined. The methodology is then applied to simulation data generated from a simple structural model subjected to a harmonic excitation. The model consists of multiple breakable springs, some of which may suffer irreversible damage when the response exceeds a threshold value or the number of cycles of motion is accumulated beyond their fatigue life. In cases of either abrupt or accumulative damages, occurrence of damage and the moment when it occurs can be clearly determined in the details of the wavelet decomposition of these data. Similar results are observed for the real acceleration data of the seismic response recorded on the roof of a building during the 1971 San Fernando earthquake. Effects of noise intensity and damage severity are investigated and presented by a detectability map. Results show the great promise of the wavelet approach for damage detection and structural health monitoring.

An artificial neural network methodology for damage detection: Demonstration on an operating wind turbine blade

Background/Objective: Early detection of damage in structures can prevent loss of lives and extend their service life. This has drawn considerable attention of researchers to applying Structural Health Monitoring (SHM) to detect structural damage. Wavelet Transform (WT) is a tool utilized to identify damage in structures by using structural responses. In this study, a WT damage detection method using Operational Deflection Shape (ODS) difference is presented. Method: A two-dimensional Continuous Wavelet Transform (CWT) is utilized to decompose the difference between the ODS of the intact and damaged structure to detect and locate damage. To demonstrate the ability of this technique, a numerical model of a steel plate is applied. Findings: The results show that the location of damage influences damage detectability when the decomposed signal is noisy. Improvements: This method eliminates the problem of border distortion experienced when conventional ODS is applied. The methodology shows the proposed techniques’ simplicity while the results show accuracy.Keywords: Cantilever Plate, Continuous Wavelet Transform, Damage Detection, Damage Severity Tree, Deflection Shape

This thesis is devoted to present some novel techniques in Structural Health Monitoring (SHM). SHM is a developing field that tries to monitor structures to make sure that they remain in their desired condition to avoid any catastrophe. SHM includes different levels from damage detection area to prognosis field. This work is dedicated to the first level, which might be considered the main and most important level. New techniques presented in this work are based on different statistical and signal processing methods such as Principal Component Analysis and its robust counterpart, Wavelet Transform, Fuzzy similarity, Andrew plots, etc. These techniques are applied on the propagated waves that are activated and captured in the structure using appropriate transducers. Piezoceramic (PZT) devices are chosen in this work to capture the signals due to their special characteristics such as high performance, low energy consumption and reasonable price. To guarantee the efficiency of the suggested techniques, they are tested on different laboratory and real scale test benchmarks, such as aluminum and composite plates, fuselage, wing skeleton, tube, etc. Because of the variety of tested benchmarks, this thesis is called damage detection in smart structures. This variety may promise the ability and capability of the proposed methods on different fields such as aerospace and gas/oil industry. In addition to the normal laboratory conditions, it is shown in this work that environmental changes can affect the performance of the damage detection and wave propagation significantly. As such, there is a vital need to consider their effect. In this work, temperature change is chosen as it is one of the main environmental fluctuation factors. To scrutinize its effect on damage detection, first, the effect of temperature is considered on wave propagation and then all the proposed methods are tested to check whether they are sensitive to temperature change or not. Finally, a temperature compensation method is applied to ensure that the proposed methods are stable and robust even when structures are subjected to variant environmental conditions. La presente tesis doctoral se dedica a la exploración y presentación de técnicas novedosas para la Monitorización y detección de defectos en estructuras (Structural Health Monitoring -SHM-) SHM es un campo actualmente en desarrollo que pretende asegurarse que las estructuras permanecen en su condición deseada para evitar cualquier catástrofe. En SHM se presentan diferentes niveles de diagnóstico, Este trabajo se concentra en el primer nivel, que se considera el más importante, la detección de los defectos. Las nuevas técnicas presentadas en esta tesis se basan en diferentes métodos estadísticos y de procesamiento de señales tales como el Análisis de Componentes Princpales (PCA) y sus variaciones robustas, Transformada wavelets, lógica difusa, gráficas de Andrew, etc. Estas técnicas de aplican sobre las ondas de vibración que se generan y se miden en la estructura utilizando trasductores apropiados. Dispositivos piezocerámicos (PZT's) se han escogido para este trabajo ya que presentan características especiales tales como: alto rendimiento, bajo consumo de energia y bajo costo. Para garantizar la eficacia de la metodología propuesta,se ha validado en diferentes laboratorios y estructuras a escala real: placas de aluminio y de material compuesto, fuselage de un avión, revestimiento del ala de un avóin, tubería, etc. Debido a la gran variedad de estructuras utilizadas, su aplicación en la industria aeroespacial y/o petrolera es prometedora. Por otra parte, los cambios ambientales pueden afectar al rendimiento de la detección de daños y propagación de la onda significativamente . En este trabajo , se estudia el efecto de las variaciones de temperatura ya que es uno de los principales factores de fluctuación del medio ambiente . Para examinar su efecto en la detección de daños, en primer lugar, todos los métodos propuestos se prueban para comprobar si son sensibles a los cambios de temperatura o no. Finalmente , se aplica un método de compensación de temperatura para garantizar que los métodos propuestos son estables y robustos incluso cuando las estructuras se someten a condiciones ambientales variantes

Data anomaly detection plays a pivotal role in structural health monitoring (SHM), enabling reliable and accurate assessment of structural condition. This paper proposes an unsupervised anomaly detection method based on the spatio-temporal sparse integration framework (STSIF), which aims to effectively distinguish various types of anomalous patterns in practical applications. First, the anomaly patterns are classified into three-level categories—normal, mild anomaly, and significant anomaly—and then a three-stage unsupervised anomaly detection framework is designed for these categories. At each stage, the characteristics-based detection indicators are proposed to detect significant, mild, and local anomaly progressively. For significant anomaly, the spatial-sparse fusion indicator is proposed and clustered with the unsupervised Gaussian mixture model (GMM). For mild anomaly, the sparse-based reconstruction detection method is proposed for qualitative analysis of noise levels, in which the Pearson correlation between the original and sparse-based reconstructed signal is computed for the GMM clustering. Overall accuracy of 98.36% and 95.51% is obtained for the data sets on two real-world bridges, respectively. Finally, regarding local anomaly detection, the proposed mask-based ResNet18 effectively emphasizes the local contextual information within the signal and substantially reduces the risk of overgeneralization. In summary, the proposed STSIF could efficiently detect various anomaly patterns without substantial training data or complex parameter tuning, providing a comprehensive detection paradigm for SHM.

This paper presents an effective method of FE model updating for health monitoring of structures by applying ambient vibration. And this method is experimented through damage detection and proved to be valid. Experiment about ambient vibration is performed on cantilever beam, and the dynamic characteristics are analyzed by NExT and ERA. The results of such experiments are compared to those of FE analysis, and this comparison enables us to overcome some errors in experiments and analysis. On the basis of improved results by the comparison, model updating is performed in order to construct a basic structure for health monitoring. For model updating, we employ direct matrix updating method (DMUM) and Error matrix method (EMM) in which ambient vibration is easily applied. The model updating by the methods are again evaluated in terms of error ratio of natural frequency, comparing each result before and after updating. Finally, we perform experiments on damage detection to verify the method of updating presented here, and evaluate its performance by eigen-parameter change method. The evaluation proves that the method of FE model updating using ambient vibration is effective for health monitoring of structure, and some further application of this method is suggested.

Chemical pressure storage tanks are containers designed to store fluids at high pressures, i.e., their internal pressure is higher than the atmospheric pressure. They can come in various shapes and sizes, and may be fabricated from a variety of materials. As aggressive chemical agents stored under elevated pressures can cause significant damage to both people and the environment, it is essential to develop systems for the early damage detection and the monitoring of structural integrity of such vessels. The development of early damage detection and condition monitoring systems could also help to reduce the maintenance costs associated with periodic inspections of the structure and unforeseen operational breaks due to unmonitored damage development. It could also reduce the related environmental burden. In this paper, we consider a hybrid material composed of glass-fiber-reinforced polymers (GFRPs) and a polyethylene (PE) layer that is suitable for pressurized chemical storage tank manufacturing. GFRPs are used for the outer layer of the tank structure and provides the dominant part of the construction stiffness, while the PE layer is used for protection against the stored chemical medium. The considered damage scenarios include simulated cracks and an erosion of the inner PE layer, as these can be early signs of structural damage leading to the leakage of hazardous liquids, which could compromise safety and, possibly, harm the environment. For damage detection, PZT sensors were selected due to their widely recognized applicability for the purpose of structural health monitoring. For sensor installation, it was assumed that only the outer GFRP layer was available as otherwise sensors could be affected by the stored chemical agent. The main focus of this paper is to verify whether elastic waves excited by PZT sensors, which are installed on the outer GFRP layer, can penetrate the GFRP and PE interface and can be used to detect damage occurring in the inner PE layer. The efficiency of different signal characteristics used for structure evaluation is compared for various frequencies and durations of the excitation signal as well as feasibility of PZT sensor application for passive acquisition of acoustic emission signals is verified.

Under the background of intelligent times based on information technology, the analysis and replacement of outliers become particularly critical for databases in the face of massive sequential data streams. In order to improve the effectiveness and practicability of the detection method, this paper determines the abnormal scores of data points in a specific data set by using the K-Sigma algorithm of the Monte Carlo method, adaptively adjusts the k value according to the abnormal scores, and marks the abnormal points.The advantages of the traditional k-sigma algorithm are fast operation speed and theoretical basis, while the disadvantages are that each index dimension is determined independently. However, in industrial production, due to the multi-index dimension of time series data, multiple size indicators of a data set are related to each other. Therefore, it is necessary to comprehensively consider whether an anomaly is abnormal by combining it with the data of other indicators. In addition, when there are no outliers in the data set, the traditional k-sigma algorithm is used to take the mean and variance of the data set as parameters. Due to the limitations of the data set itself, some data points will be mislabeled as outliers. Through the k-sigma algorithm based on the Monte Carlo method, we can effectively solve the above problems.The k-sigma algorithm based on the Monte Carlo method can generate a normal distribution according to the distribution of original data points, and extract a large quantity of data from the distribution to generate Monte Carlo data set. The Mahalanobis distance li from each sample point to the mean in the Monte Carlo data set is calculated and compared with the Mahalanobis distance lj from the samples to be detected to the mean in the original data set. According to the number of sample points satisfying li < lj in the Monte Carlo data set, the value of the parameter kj is determined adaptively, and thus the outliers are determined. We implemented the k-sigma algorithm based on the Monte Carlo method through python, evaluated the effectiveness of the algorithm by accuracy, recall rate, and F1 score, and compared it with some machine learning algorithms. The results verified the feasibility and effectiveness of the algorithm, which can be used for real-time anomaly detection in the energy management database.

Fumaric acid esters are considered efficacious and safe drugs for the treatment of psoriasis. Renal damage, caused either by acute renal injury or Fanconi syndrome, is a recognized side-effect of this therapy. To investigate whether the measurement of urinary excretion of β2-microglobulin, a marker of renal proximal tubular dysfunction, allows early detection of kidney damage before an increase in serum creatinine or significant proteinuria occurs. Urinary β2-microglobulin excretion was measured regularly in 23 patients undergoing fumaric acid ester therapy. Urinary β2-microglobulin remained normal in all 10 male patients. Three (23%) out of 13 female patients experienced an increase in urinary β2-microglobulin excretion. In two of these patients a sharp increase was observed in association with high doses. One further patient had moderately elevated levels on rather low doses of fumaric acid esters. After discontinuing treatment, urinary β2-microglobulin levels returned to normal within a few weeks. Determination of urinary β2-microglobulin possibly allows early detection of renal damage by fumaric acid esters. Female patients seem to be prone to this side-effect, especially when taking high doses.

Air quality has always been a hot issue of concern to the people, the environmental protection department and the government. Among the massive air quality data, abnormal data can interfere with subsequent experiments and analysis. Therefore, it is necessary to detect abnormal data to improve the accuracy of the data. However, traditional air outlier detection methods require at least one year's data to make inferences about air quality. This paper firstly analyzes the characteristics of air quality big data, and then proposes a framework based on Bayesian non-parametric clustering, namely Dirichlet Process (DP) clustering framework, to realize the outlier detection of air quality. The framework optimizes Gaussian mixture model into infinite Gaussian mixture model according to the results of data analysis, and uses neural network to cluster the data processed by infinite Gaussian mixture model, which effectively improves the clustering accuracy and avoids the need of collecting a large number of training data.

The application of vibration-based Structural Health Monitoring (SHM) for damage detection is characterised by three fundamental aspects: the features extracted as representative of the structural condition that can be directly linked to some form of damage, the metrics selected as novelty or damage index, and the statistical model or classifier built to identify underlying patterns indicative of changes in the structure’s state. Focusing on the first step to improve the performance of vibration-based SHM strategies, the extracted features should be robust to noise, sensitive to the presence of a specific type of damage. Further, damage should induce patterns that are distinguishable from environmental and operational variabilities and other forms of damage such as ageing phenomena. In this paper, the problem is framed as an outlier detection problem and the the use of different modal parameters as Damage Sensitive Features (DSFs) is investigated, evaluating them based on the detection performance of an unsupervised One-Class Support Vector Machine (OCSVM) classifier. In particular, an artificial dataset is generated from the calibrated numerical model of a three-storey steel frame structure in different damage scenarios. The methodology is validated against available experimental data. For the case investigated the natural frequencies were sensitive to damage and robust to noise.

Real-time anomaly detection based on long short-Term memory and Gaussian Mixture Model