Damage Detection Of Civil Structures Research Articles

Research on Damage Detection of Civil Structures Based on Machine Learning of Multiple Vegetation Index Time Series

Research on Damage Detection of Civil Structures Based on Machine Learning of Multiple Vegetation Index Time Series

Time Series Analysis Methodology for Damage Detection in Civil Structures

Structural health monitoring (SHM) methodologies employing data-driven techniques are becoming increasingly popular for detection of structural damage at the earliest stage possible. With measured vibration signals from the structure, time series modeling methods provide quantitative means for extracting such features that can be utilized for damage diagnosis. In this study, one-step prediction error of an autoregressive (AR) model over a data set is used as damage indicator. In particular, the difference between the prediction of the AR model that is fit to the measured acceleration signal obtained from the intact structure and actual measured signals collected for different damage states of the structure are interrogated for diagnosis purposes. More specifically, the standard deviation of the residual error is employed to locate the damaged region. Singular-value decomposition (SVD) is employed to find the optimal order for an AR model created using the impulse responses of the system. Numerical simulations are carried out using the impulse responses acquired from a four-story frame structure contaminated with additive noise including single and multiple damaged elements. The results of the simulations demonstrate that the method can be effectively employed to detect and locate damage. The performance of the proposed procedure are further demonstrated using the impact data acquired from a reinforced concrete frame for real applications.

VIBRATION-BASED DAMAGE DETECTION FOR ONE-STORY STEEL FRAME STRUCTURE USING MODE SHAPE CURVATURE

Structural health monitoring techniques, particularly vibration-based damage detection, have gained significance in assessing civil structure condition. This paper focuses on utilising mode shape curvature for damage detection in a one-story steel frame structure. The study aims to overcome traditional inspection limitations by exploring vibration-based approaches. Experimental investigation is conducted to analyse intact and damaged structural modal behaviour. Modal analysis technique extracts modal frequencies and mode shapes, enabling analysis of mode shape curvature for damage detection and localisation. Preliminary findings show that damaged structures display deviations in mode shapes and reduced natural frequencies, providing evidence of structural damage. However, a significant issue arises near the support, where unexpected patterns emerge in the Total Damage Index (TDI) with increasing damage severity. This finding challenges the expected correlation between severity levels and TDI values, highlighting the need to consider factors like fixed supports. Misleading signs of damage in some segments underscore the importance of cautious result interpretation and accounting for noise. Future studies should focus on noise resistance, false indication mitigation, and understanding segments with fixed supports to enhance mode shape curvature analysis’s reliability for damage detection in civil structures.

Transfer Learning for CNN-Based Damage Detection in Civil Structures with Insufficient Data

Among various methods proposed for health monitoring of structures, deep learning-based techniques with their powerful performance have attracted considerable attention in recent years. However, a major problem with these methods is that they usually need large amounts of data in the training phase, while such data may not be available in real applications. In this study, compact one-dimensional (1D) convolutional neural networks (CNNs) are utilized that require less data for training. The study is comprised of two parts: the first stage aims to develop a compact CNN that can recognize damages in a structure with high accuracy, when data are provided to some extent. The problem of inadequate training data in health monitoring of experimental and real-life structures is then investigated in the second part. Transfer learning is used to deal with this problem. A compact CNN is utilized as the source domain network and the target domain network receives all of its knowledge from this source. Acceleration time histories from a numerical model, an experimental structure, and a full-scale bridge are utilized to validate the proposed methodology. According to the results, the compact CNN can reach 100% accuracy when data are available for training. Also, for the case of insufficient data, using a compact network as well as transfer learning causes considerable improvement (about 95%) in the accuracy of damage detection.

Statistical analysis of the modal properties of a seismically-damaged five-story RC building identified using ambient vibration data

Damage detection (DD) is one of the primary goals of health monitoring of civil structures. Vibration-based techniques aim to identify the modal properties (or vibration characteristics) of structures and are one of the most popular approaches for structural damage detection. The use of vibration characteristics (natural frequencies, damping ratios, and mode shapes) for DD purposes is based on the premise that these characteristics change when the structure suffers damage, since the modal properties depend on the physical properties (mass, stiffness, and damping) of the structure. The use of output-only measurements (e.g., ambient vibration or AV) is the most popular approach for damage detection of civil structures. AV data recorded before and after the structure has potentially suffered damage can be used for DD. However, application of vibration-based DD using AV requires an accurate and reliable estimation of the modal properties and their variability (or uncertainty) in order to genuinely determine the existence of damage. This study presents a comprehensive statistical analysis of the identified modal properties of a full-scale five-story reinforced concrete building using AV data. The building specimen was tested on the NHERI@UCSD shake table in base-isolated and fixed-base configurations. On April 6, 2012, about two weeks before the start of the seismic tests in the base-isolated configuration, a dense array of twenty accelerometers was deployed on the structure and AV data were recorded continuously until May 18, 2012, three days after completion of the seismic tests in the fixed-base configuration. In its fixed-base configuration, the building was subjected to a sequence of six earthquake motions that progressively damaged the structure. Two popular methods of operational modal analysis are used to automatically identify the modal properties of the fixed-base building at different damage states using the recorded AV data. A statistical analysis of the identified modal parameters is performed to investigate the statistical variability and accuracy of the system identification results. The variability of the identified modal parameters due to environmental conditions is also investigated.

Damage Detection in Largely Unobserved Structures under Varying Environmental Conditions: An AutoRegressive Spectrum and Multi-Level Machine Learning Methodology.

Vibration-based damage detection in civil structures using data-driven methods requires sufficient vibration responses acquired with a sensor network. Due to technical and economic reasons, it is not always possible to deploy a large number of sensors. This limitation may lead to partial information being handled for damage detection purposes, under environmental variability. To address this challenge, this article proposes an innovative multi-level machine learning method by employing the autoregressive spectrum as the main damage-sensitive feature. The proposed method consists of three levels: (i) distance calculation by the log-spectral distance, to increase damage detectability and generate distance-based training and test samples; (ii) feature normalization by an improved factor analysis, to remove environmental variations; and (iii) decision-making for damage localization by means of the Jensen–Shannon divergence. The major contributions of this research are represented by the development of the aforementioned multi-level machine learning method, and by the proposal of the new factor analysis for feature normalization. Limited vibration datasets relevant to a truss structure and consisting of acceleration time histories induced by shaker excitation in a passive system, have been used to validate the proposed method and to compare it with alternate, state-of-the-art strategies.

A Literature Review on the Development of Remote Sensing in Damage Detection of Civil Structures

Remote sensing technologies have a direct impact on gaining structural damage information due to their powerful flexibilities, such as wide field of view, non-contact, low cost, and fast response capacities. It is because remote sensing is often applied to monitor near-real-time damage for large-scale events. Therefore, diverse types of remote sensing data became available and various methods have been designed and reported for structural damage assessment. In this line, a number of remote sensing procedures have been proposed to develop an extensive range of temporal, spectral, and spatial parameters. In this study, a comparative review is conducted in order to present the applied remote sensing-based damage detection approaches in buildings and bridges. It should be noted that the survey is supported by an extensive list for up-to-date references. Based on this review, it can be concluded that remote sensing has widely attracted attentions in different structural engineering fields due to its ability in providing fast response in terms of continuous monitoring for large areas after a natural hazard.

Unsupervised Structural Damage Detection Technique Based on a Deep Convolutional Autoencoder

Structural health monitoring (SHM) is a hot research topic with the main purpose of damage detection in a structure and assessing its health state. The major focus of SHM studies in recent years has been on developing vibration‐based damage detection algorithms and using machine learning, especially deep learning‐based approaches. Most of the deep learning‐based methods proposed for damage detection in civil structures are based on supervised algorithms that require data from the healthy state and different damaged states of the structure in the training phase. As it is not usually possible to collect data from damaged states of a large civil structure, using such algorithms for these structures may be impractical. This paper proposes a new unsupervised deep learning‐based method for structural damage detection based on convolutional autoencoders (CAEs). The main objective of the proposed method is to identify and quantify structural damage using a CAE network that employs raw vibration signals from the structure and is trained by the signals solely acquired from the healthy state of the structure. The CAE is chosen to take advantage of high feature extraction capability of convolution layers and at the same time use the advantages of an autoencoder as an unsupervised algorithm that does not need data from damaged states in the training phase. Applications on the two numerical models of IASC‐ASCE benchmark structure and a grid structure located at the University of Central Florida, as well as the full‐scale Tianjin Yonghe Bridge, prove the efficiency of the proposed algorithm in assessing the global health state of the structures and quantifying the damage.

Multifractal analysis and wavelet leaders for structural damage detection of structures subjected to earthquake excitation

This work is an effort to join, for the first-time, multifractal analysis and damage detection in civil structures subjected to strong ground seismic motions. Specifically, based on the singularity spectrum quantitative and qualitative criteria are proposed. The qualitative criteria are based on the concave of singularity spectrum of damage and undamaged structure. The proposed quantitative criterion is based on calculation of damage index taken the parameters of singularity spectrum. In order to achieve the above goal, a robust signal processing method, which is known as multifractal wavelet leader (MFWL) is used. The multifractal analysis is a tool to calculate fractal properties as well as scaling behavior of the structural response excited by an earthquake. The singularity spectrum is obtained from the Legendre-transformation to Holder exponents. In this paper, a parameter which is based on the shape of singularity spectrum and can identify the damage in the structure is proposed. The proposed method is an output-only approach for damage detection. Considering that the dynamic behavior of an inelastic system subjected to strong ground motion appears to be a non-stationary process, the above procedure of multifractal wavelet leader is suitable to retrieve the simulation response data. The findings from the analysis show that the MFWL is an appropriate scheme for structural damage detection.

Improved synchronization-based approach for online damage detection of civil structures

In this paper, a new online damage detection method is developed using displacement and velocity measurements of the structure. The proposed method is an improved synchronization-based approach in which any changes in system parameters can be measured by synchronizing a mathematical model with the real model of the structure in real time. Several damage scenarios for structures subjected to different seismic ground motion acceleration records are studied to validate the effectiveness of the proposed method. The effect of noise on the performance of the proposed algorithm is also investigated and discussed. Numerical results demonstrate that the proposed technique has the capability of tracking the abrupt changes of stiffness and damping parameters with high accuracy in real time.

Separating damage from environmental effects affecting civil structures for near real-time damage detection

Analyzing changes in vibration properties (e.g. natural frequencies) of structures as a result of damage has been heavily used by researchers for damage detection of civil structures. These changes, however, are not only caused by damage of the structural components, but they are also affected by the varying environmental conditions the structures are faced with, such as the temperature change, which limits the use of most damage detection methods presented in the literature that did not account for these effects. In this article, a damage detection method capable of distinguishing between the effects of damage and of the changing environmental conditions affecting damage sensitivity features is proposed. This method eliminates the need to form the baseline of the undamaged structure using damage sensitivity features obtained from a wide range of environmental conditions, as conventionally has been done, and utilizes features from two extreme and opposite environmental conditions as baselines. To allow near real-time monitoring, subsequent measurements are added one at a time to the baseline to create new data sets. Principal component analysis is then introduced for processing each data set so that patterns can be extracted and damage can be distinguished from environmental effects. The proposed method is tested using a two-dimensional truss structure and validated using measurements from the Z24 Bridge which was monitored for nearly a year, with damage scenarios applied to it near the end of the monitoring period. The results demonstrate the robustness of the proposed method for damage detection under changing environmental conditions. The method also works despite the nonlinear effects produced by environmental conditions on damage sensitivity features. Moreover, since each measurement is allowed to be analyzed one at a time, near real-time monitoring is possible. Damage progression can also be given from the method which makes it advantageous for damage evolution monitoring.

A Bayesian state-space approach for damage detection and classification

The problem of automatic damage detection in civil structures is complex and requires a system that can interpret collected sensor data into meaningful information. We apply our recently developed switching Bayesian model for dependency analysis to the problems of damage detection and classification. The model relies on a state-space approach that accounts for noisy measurement processes and missing data, which also infers the statistical temporal dependency between measurement locations signifying the potential flow of information within the structure. A Gibbs sampling algorithm is used to simultaneously infer the latent states, parameters of the state dynamics, the dependence graph, and any changes in behavior. By employing a fully Bayesian approach, we are able to characterize uncertainty in these variables via their posterior distribution and provide probabilistic estimates of the occurrence of damage or a specific damage scenario. We also implement a single class classification method which is more realistic for most real world situations where training data for a damaged structure is not available. We demonstrate the methodology with experimental test data from a laboratory model structure and accelerometer data from a real world structure during different environmental and excitation conditions.

Damage Detection of Structures Subject to Nonlinear Effects of Changing Environmental Conditions

Damage detection of civil structures has been carried out by mainly analysing the vibration properties of the structures which change when damages occur. However, these properties are also affected by the changing environmental conditions the structures are face with, and these conditions usually produce nonlinear effects on the vibration properties. Hence, a method is proposed in this paper to analyse structures subjected to nonlinear effects of environmental conditions. The method first applies Principal Component Analysis (PCA) on a bank of damage sensitivity features, followed by applying Gaussian Mixture Model on the obtained first principal component scores to cluster the data into several linear regions. By creating a baseline for each linear region using two extreme and opposite environmental conditions, and adding new measurements to the baseline one at a time followed by applying PCA, damage detection can be achieved. The method is validated on a numerical truss structure model and on the Z24 Bridge. The results demonstrate the ability of the method to analyse structures under nonlinear environmental effects.

Analyses of the most influential factors for vibration monitoring of planetary power transmissions in pellet mills by adaptive neuro-fuzzy technique

Vibration-based structural health monitoring is widely recognized as an attractive strategy for early damage detection in civil structures. Vibration monitoring and prediction is important for any system since it can save many unpredictable behaviors of the system. If the vibration monitoring is properly managed, that can ensure economic and safe operations. Potentials for further improvement of vibration monitoring lie in the improvement of current control strategies. One of the options is the introduction of model predictive control. Multistep ahead predictive models of vibration are a starting point for creating a successful model predictive strategy. For the purpose of this article, predictive models of are created for vibration monitoring of planetary power transmissions in pellet mills. The models were developed using the novel method based on ANFIS (adaptive neuro fuzzy inference system). The aim of this study is to investigate the potential of ANFIS for selecting the most relevant variables for predictive models of vibration monitoring of pellet mills power transmission. The vibration data are collected by PIC (Programmable Interface Controller) microcontrollers. The goal of the predictive vibration monitoring of planetary power transmissions in pellet mills is to indicate deterioration in the vibration of the power transmissions before the actual failure occurs. The ANFIS process for variable selection was implemented in order to detect the predominant variables affecting the prediction of vibration monitoring. It was also used to select the minimal input subset of variables from the initial set of input variables – current and lagged variables (up to 11 steps) of vibration. The obtained results could be used for simplification of predictive methods so as to avoid multiple input variables. It was preferable to used models with less inputs because of overfitting between training and testing data. While the obtained results are promising, further work is required in order to get results that could be directly applied in practice.

Moving-window extended Kalman filter for structural damage detection with unknown process and measurement noises

The extended Kalman filter (EKF), as a popular tool for optimally estimating system state from noisy measurement, has been used successfully in various areas over the past several decades. However, classical EKF has several limitations when applied to structural system identification; thus, researchers have proposed a number of variations for this method. The current study focuses on using EKF for real-time system identification and damage detection in civil structures. An improved EKF approach, called moving-window EKF (MWEKF), is proposed in this paper after a discussion on the problems associated with the application of classical EKF in time-variant systems. The proposed approach uses the moving-window technique to estimate several statistical properties. MWEKF is more robust and adaptive in structural damage detection compared with classical EKF because of the following reasons: (1) it is insensitive to the selection of the initial state vector; (2) it exhibits more accurate system parameter identification; and (3) it is immune to the inaccurate assumption of noise levels because measurement and process noise levels are estimated in this approach. The salient features of MWEKF are illustrated through numerical simulations of time-variant structural systems and an experiment on a three-story steel shear building model. Results demonstrate that MWEKF is a robust and effective tool for system identification and damage detection in civil structures.

Autoregressive statistical pattern recognition algorithms for damage detection in civil structures

Statistical pattern recognition has recently emerged as a promising set of complementary methods to system identification for automatic structural damage assessment. Its essence is to use well-known concepts in statistics for boundary definition of different pattern classes, such as those for damaged and undamaged structures. In this paper, several statistical pattern recognition algorithms using autoregressive models, including statistical control charts and hypothesis testing, are reviewed as potentially competitive damage detection techniques. To enhance the performance of statistical methods, new feature extraction techniques using model spectra and residual autocorrelation, together with resampling-based threshold construction methods, are proposed. Subsequently, simulated acceleration data from a multi degree-of-freedom system is generated to test and compare the efficiency of the existing and proposed algorithms. Data from laboratory experiments conducted on a truss and a large-scale bridge slab model are then used to further validate the damage detection methods and demonstrate the superior performance of proposed algorithms.

Monitoreo de Salud Estructural

Non-destructive methods for damage detection in civil structures have recently gained the attention of researchers throughout the world due to the advances in instrumentation technology and digital signal processing. Structural Health Monitoring techniques allows the presence, location and magnitude of damage to be detected due to changes in the static and dynamic characteristics of the structure. These techniques can be implemented remotely and in real-time, which can significantly reduce inspection costs and provide an instantaneous alert signal due to damage in the structure. The School of Civil Engineering and Geomatics of the Universidad del Valle has a Structural Health Monitoring research project that includes real-time monitoring of the Hormiguero bridge and an agreement with the Universidad Tecnológica de Pereira for joint monitoring of the Pereira-Dos Quebradas Viaduct.

Experimental Investigation on Statistical Moment-based Structural Damage Detection Method

Although vibration-based structural damage detection methods have demonstrated various degrees of success, the damage detection of civil structures still remains as a challenging task. The main obstacles include the insensitivity to local damage and the high sensitivity to measurement noise. A new structural damage detection method based on the statistical moments of dynamic responses of a structure has been recently proposed by the authors, and the numerical study manifested that the proposed method is sensitive to local structural damage but insensitive to measurement noise. The experimental investigation on this method is presented in this article. Three shear building models with and without damage were built and subjected to ground motions generated by a shaking table. The displacement and acceleration responses of each building model at each floor were recorded. The recorded ground motion and building responses as well as identified structural damping ratios were then used to identify damage locations and severities using the statistical moment-based damage detection method. The identified damage locations and severities were compared with the theoretical values. The comparison is found satisfactory, and the method proposed is effective and feasible.

Piezoelectric sensor based nondestructive active monitoring of strength gain inconcrete

The recent advent of smart materials, such as piezoelectric materials, shape-memory alloys,and optical fibers, has added a new dimension to present structural health monitoringtechniques. In particular, the electro-mechanical impedance (EMI) sensing techniqueutilizing piezoelectric materials has emerged as a potential tool for the implementation of abuilt-in monitoring system for damage detection of civil structures. However, there is littleeffort to apply this technique for concrete monitoring. In this study, an effort to extend theapplicability of the EMI sensing technique is made for strength gain monitoringof early age concrete. PZT (piezoelectric lead zirconate titanate) patches areemployed to sense the EMI signature of curing concrete. A series of experiments wasconducted on concrete specimens to verify the applicability of the EMI sensingtechnique. The results show the excellent potential of the EMI sensing techniqueas a practical and reliable nondestructive method for strength gain monitoring.

Wavelet Decomposition-Based Approach for Fast Damage Detection of Civil Structures

This paper presents a method, which can be implemented under limited resources such as wireless sensor network devices, for damage detection of civil structures. The detection method is applied to transient waves from vibration responses, and it consists of Haar wavelet decomposition, thresholding, quantization, and differentiating metric. Before the differentiating metric, a target signal is compressed into the wavelet coefficients .signature.. The results indicate that a single measurement channel without time synchronization of other measurement points is effective for instant damage detection through the case study using the ASCE structural health monitoring (SHM) benchmark program. It is also shown that we can reduce the data to a small fraction of the original data volume using this method, while retaining enough information to detect structural damage and its severity. The approach illustrated in this paper implies a promise of wavelet-based data compression and analysis for the increasing volume of SHM data.