Structural Damage Quantification Research Articles

Structural damage quantification using long short-term memory (LSTM) auto-encoder and impulse response functions

This paper presents an approach for structural damage quantification using a long short-term memory (LSTM) auto-encoder and impulse response functions (IRF). Among time domain responses-based methods for structural damage identification, using IRF is advantageous over the original time domain responses, since IRF consists of information of system properties and is loading effect independent. In this study, IRFs are extracted from the acceleration responses measured from different locations of structures under impact force excitations. The obtained IRFs are concatenated. Moving averaging with a suitable window size is performed to reduce random variations in the concatenated responses. Further, principal component analysis is performed for dimensionality reduction. These selected principal components are then fed to the LSTM auto-encoder for structural damage identification. A noise layer is added as an input layer to the LSTM auto-encoder to regularise the model. The proposed model consists of two phases: (1) reconstruction of the selected “principal components” to extract the features; and (2) damage identification of structural elements. Numerical studies are conducted to verify the accuracy of the proposed approach. The results demonstrate that the proposed approach can accurately identify and quantify structural damage for both single- and multiple-element damage cases with noisy measurements, as well as uncertainties in the stiffness parameters. Furthermore, the performance of the proposed approach is evaluated using the limited measurements from a few sensors.

A hybrid multiple input multiple output (MIMO) model for simultaneous localization and quantification of structural damage in beam structures

Minor cracks are common types of damage in beam-like structures that must be detected early to prevent complete structural failure. Detecting and addressing such damage is crucial for maintaining structural integrity, extending service life, and minimizing operational costs. In recent years, various methods for quantifying and localizing damage using vibration measurements have been developed. However, many of these methods struggle to quantify and accurately localize the damage simultaneously. This paper addresses this challenge by applying novel hybrid multi-input and multi-output (MIMO) regression algorithms. The proposed approach combines modal analysis data, which includes frequencies, with machine learning algorithms to achieve simultaneous damage localization and quantification in damaged beams. Both numerical and experimental modal analyses have been conducted to determine the first six modal frequencies. Multiple hybrid MIMO models have been trained and tested using the frequencies obtained. The trained models have been used to estimate damage location and severity, and the results are compared. The models’ performance was evaluated using seven key metrics, including mean squared error (MSE), coefficient of determination (R2), mean absolute error (MAE), and others. The Random Forest (RF) model demonstrated exceptional performance by achieving a significantly higher R2 value of 0.997 for the precise prediction of both damage location and magnitude. Optimizing these models through hyperparameter tuning significantly improved their accuracy and reliability. This approach exhibits superior efficacy, surpassing existing models, and holds promise for advancing structural health monitoring in critical applications.

Time‐varying reliability analysis based on hybrid Kalman filtering and probability density evolution

AbstractThe study introduces a novel approach for time‐varying reliability analysis of structures called “hybrid UKF‐PDEM” by integrating the unscented Kalman filter (UKF) and the probability density evolution method (PDEM). The UKF estimates the displacement, velocity, stiffness, and damping parameters of a structure at each time step subjected to dynamic loading for structural damage quantification. The estimated parameters at each time step are then input into the PDEM to calculate the time‐varying probability density function (PDF) of the estimated states. The estimated PDF is used to update the uncertainty matrix of the estimated states in each iteration and to determine the time‐varying reliability curves of the structure. To demonstrate the effectiveness of the proposed method, we applied it to a numerical model of a three‐degree‐of‐freedom system and a full‐scale seven‐story building with different damage scenarios. The method is used to estimate the level of damage and calculate the corresponding reliability curve of the system over time for each damage scenario, utilizing the estimated structural responses and stiffness values. The extracted reliability values for each damage scenario follow the level of damage over time. This study shows that the newly developed method is computationally efficient for building a digital twin and enables real‐time damage identification and reliability analysis in various structural systems. The method's applicability to different types of structures highlights its versatility and potential for widespread use in assessing the integrity of buildings and infrastructure.

Detection, Localisation and Quantification of Structural Damage Using Changes in Modal Characteristics

The article defines and explains methods for detecting and locating damage and quantifying its extent on an example of a finite element model of a simple beam. The defined methods are based on a comparison of the decreases in the bending natural frequencies of the nominally damaged models against the intact model. This comparison assigns natural frequency decreases to curvature of a specific mode shape, which allows creation of an analytical reference model using polynomial regression, which assigns relative natural frequency decreases to the position on the beam. The localisation method assumes that the vector of the relative natural frequency decrease ratios are unique for each position on the beam. The quantification method considers as a relevant comparative quantity, slopes of relative natural frequency decreases, as a function of damage extent. Examples of damage localisation and quantification are defined, and the relative errors of these methods are analysed.

Model-informed deep learning strategy with vision measurement for damage identification of truss structures

Structural damage identification approaches can be divided into two categories, i.e. data-driven approaches via statistical pattern recognition and model-based approaches via finite element (FE) model updating. These two approaches have their own merits, and their main shortcomings can be remedied by each other’s merits. Therefore, this study proposed a deep learning-based damage identification strategy involving both data-driven and model-based approaches, termed as model-informed deep learning (MIDL)-based strategy. This strategy first proposes a vision-based displacement estimation approach to extract structural displacement responses from video data. This approach reduces the displacement drift induced by conventional optical flow approaches and improves the tracking accuracy of feature points. Then, a calibrated FE model is built to construct data sets with different damage levels via FE model updating and time-history analysis. Following this, a one-dimensional convolutional neural network (1D CNN) is established to detect and localize structural damage by using direct displacement responses. Finally, FE model updating is performed again to quantify structural damage level with constrained targets. A truss structure is further used to evaluate the accuracy of the proposed strategy experimentally. Results illustrate that the proposed MIDL strategy, which uses time series to localize the structural damage, achieves a global location accuracy of 86.09% and avoids the feature extraction process. Meanwhile, with the known damage location, the efficiency and accuracy of structural damage quantification via rerunning model updating can also be significantly improved. In addition, the displacements estimated by the proposed approach have a good match with ground truth values with error standard deviations of less than 0.3 mm.

A chaos game Optimization-based model updating technique for structural damage identification under incomplete noisy measurements and temperature variations

Dealing with structural damage identification (SDI) considering limited noise-contaminated measurements and varying environmental conditions is still a challenge due to the difficulties regarding the localization and quantification of structural damage. If the challenging issue is not considered in SDI process, the SDI techniques may become unreliable and have limitations in their implementation in in-service structures. In this regard, the article proposes an optimization-based model updating technique for SDI taking into consideration both incomplete noisy measurements and temperature variations. First, it is assumed in the finite element model of the monitored structure that the relation between temperature conditions and structural materials is temperature-dependent to synthetically generate damage-induced vibration data under varying environmental temperatures. Then, the inverse SDI problem is recast into a constrained optimization problem and a newly developed algorithm called Chaos Game Optimization (CGO) is adopted for solving this problem with high efficiency. The efficiency of the CGO algorithm is assessed by comparison with other four parameter-less optimization algorithms. Finally, three engineering numerical examples are conducted to demonstrate the performance of the proposed technique. The results of the numerical investigations indicated that for the monitored structures made of metallic material (i.e., steel and aluminum), the proposed technique can accurately localize and quantify single and multi-damage even in incomplete and noisy modal data and environmental temperature changes. Meanwhile, for the monitored structures made of concrete material, the ambient temperature variations significantly affect the damage detectability and localizability of the proposed technique.

DETECT-AGING blind prediction contest: a benchmark for structural health monitoring of masonry buildings

The installation of monitoring systems on buildings allows analyzing variations in structural parameters over time, creating room for detection of damage. Structural Health Monitoring (SHM) systems have the potential to support pro-active risk management, where structural interventions are planned if specific thresholds related to target performance losses are achieved.DETECT-AGING is a research project of relevant national interest that was funded by the Italian Ministry of University and Research (MUR) through the PRIN 2017 programme. The project started in September 2019 and involves the universities of Bologna, Genova, Napoli Federico II, and Perugia. The main goal of the project is to develop a new analytical-instrumental approach aimed at the quantitative assessment of the effects of aging and material degradation on structural safety of cultural heritage, with special focus on masonry structures. Based on a combined use of structural models and health monitoring systems, indications and operational tools will be provided for the identification and quantification of structural damage, supporting the management of built cultural heritage. To this purpose, a two-storey masonry building, having a single room with a vault at the first floor and a timber roof, was built with the aim of being monitored and progressively and will be damaged during the project. It is equipped with a hybrid SHM system managed by the University of Perugia, which is based on both vibration and strain measurements. The present paper illustrates the main features of the case-study building and presents the results of the experimental program aimed at characterizing the mechanical properties of masonry the materials used. The final part of the paper presents a blind prediction contest based on prediction of modal features of the building in different damaged configurations.

Structural damage quantification in shear buildings using mode shape slope ratio

Mode shape-based techniques are very popular among researchers since the last two decades and varying levels of success have been achieved towards identifying damage using structural mode shapes. This study presents a novel formulation to establish the expression of damage severity in the form of mode shape slope. The derived closed-form solution directly correlates the percentage damage intensity with the derivatives of the changes in mode shapes in shear buildings, the most common class of civil structures. Numerical experimentation involving a 16-storey two-dimensional shear building has been performed to show the applicability of the proposition in damage quantification. First three mode shapes are used for this purpose. A detailed parametric study comprising of location as well as noise sensitivity has also been carried out to test the robustness of the proposition. Further, the derived expression has been applied to quantify damages in multiple locations. A relatively shorter structure of eight-storey shear building has also been considered to check the efficiency of the proposed algorithm in structural damage quantification. However, if any mode depicts a very low slope value at the damage location, that particular mode could not be used for damage quantification. Finally, an extensive experimental program on a three-dimensional six-storey miniature model of a shear building ensures its practical applicability under real measurement condition in a laboratory environment. Therefore, this novel mode shape slope-based damage quantification technique can be widely applicable to be appropriate for various circumstances.

A deep learning approach for electromechanical impedance based concrete structural damage quantification using two-dimensional convolutional neural network

Deep learning approach using convolutional neural networks (CNNs) has ushered in numerous breakthroughs in image-based recognition field, but the electromechanical impedance/admittance (EMI/EMA)-based structural damage identification by CNN remains being refined. This paper proposed a deep learning approach for the raw EMA-based rapid damage quantification on concrete structure utilizing two-dimensional (2D) CNNs. In the approach, the EMA signatures are first split into multiple sub-range responses, among which corresponding to the maximum indices namely root mean square deviations (RMSDs) are selected to construct the input of CNNs for training, and then damage severity degree could be rapidly predicted. The proposed approach is verified through crossover experiments of detecting multiple mass loss damages on a cubic concrete structure. Effect of input size on the performance of the approach is also evaluated by developing different CNN models. Experimental results confirm that the proposed approach is of high accuracy and efficiency even to tiny damages, thus paving a promising way to the real-life monitoring for concrete structures.

Structural damage quantification using ensemble‐based extremely randomised trees and impulse response functions

Structural damage quantification using ensemble‐based extremely randomised trees and impulse response functions

Gaussian process regression for active sensing probabilistic structural health monitoring: experimental assessment across multiple damage and loading scenarios

In the near future, Structural Health Monitoring (SHM) technologies will be capable of overcoming the drawbacks in the current maintenance and life-cycle management paradigms, namely, cost, increased downtime, less-than-optimal safety management paradigm and the limited applicability of fully-autonomous operations. One of the most challenging tasks is structural damage quantification. Existing quantification techniques face accuracy and/or robustness issues when it comes to varying operating and environmental conditions. In addition, the damage/no-damage paradigm of current industrial frameworks does not offer much information to maintainers on the ground for proper decision-making. In this study, a novel structural damage quantification framework is proposed based on the widely-used Damage Indices (DIs) and Gaussian Process Regression Models (GPRMs) in order to overcome the aforementioned shortcomings. The proposed framework takes a simple approach to the damage quantification problem by using DI values for training, and provides confidence bounds on the estimated states. In addition, the proposed method is shown to quantify multiple structural states simultaneously from incoming DI test points. This framework is applied to three test cases: a Carbon Fibre-Reinforced Plastic (CFRP) coupon with attached weights as simulated damage, an aluminum coupon with a notch and an aluminum coupon with attached weights as simulated damage under varying loading states. The novel state prediction method presented herein is applied to single-state quantification in the first two test cases, as well as the third one assuming the loading state is known. Finally, the proposed method is applied to the third test case assuming neither the damage size nor the load is known in order to predict both simultaneously from incoming DI test points. In applying this framework, two forms of GPRMs (standard and variational heteroscedastic) are used in order to critically assess their performance with respect to the three test cases.

Optical Coherence Tomography in Glaucoma

Optical coherence tomography (OCT) evaluates the optic nerve head (ONH), peripapillary retinal nerve fiber layer, and inner retinal layers at the macula for detecting structural abnormalities in glaucoma. The optic disc and retinal nerve fiber layer are the principal sites of apparent glaucomatous damage, which precedes glaucomatous visual field defects. The currently accepted gold standards for glaucoma diagnosis are ONH analysis for structural changes and perimetry to monitor changes in function. However, imaging technology OCT is playing an increasing role in glaucoma diagnosis, monitoring of disease progress, and quantification of structural damage. This article focused on the principle of OCT, applications of both anterior and posterior segment OCT in glaucoma management, and interpretation of OCT reports.

Assessment of Internal Damage in Sandwich Structures by Post-Processing of Mode Shapes Using Curvelet Transform.

Identification and quantification of structural damage is one of the crucial aspects of proper maintenance of mechanical and civil structures, which is directly related to their integrity and safety. The paper presents a novel approach for detecting various types of damage in sandwich structures by processing the mode shapes using a hybrid algorithm based on the curvelet transform and the standardized damage index concept. The proposed approach uses the properties of directional selectivity, absence of the boundary effect, typical of such a class of transforms, and excellent filtration capabilities of the curvelet transform as well as the classification hypothesis in the standardized damage index, which allows the exclusion of irrelevant information and emphasizes proper damage location and shape. The proposed hybrid algorithm allowed to successfully identify a subsurface core damage in sandwich structures, such as local lack of a core or its debonding from facings. The performed quantification study aimed to evaluate the correctness of identified damage shape confirmed the validity and accuracy of the proposed algorithm not only for the damage detection and localization but also for the estimation of the size of structural damage.

An advanced Wigner–Ville time–frequency analysis of Lamb wave signals based upon an autoregressive model for efficient damage inspection

It is quite straightforward to generate and acquire ultrasonic guided waves in metallic or composite structures to investigate structural defects; however, the interpretation and evaluation of the reflected/transmitted signal to extract useful information is a challenging task. This is primarily due to the dispersion and multimodal behavior of Lamb waves, which are dependent on the frequency of the excitation wave and the thickness of the material under investigation. These multimodal and dispersive behaviours lead to a complex waveform structure, and therefore require an advanced signal processing technique to decipher useful information in the time and/or frequency domains. For this purpose, due to its desirable mathematical properties, the Wigner–Ville distribution (WVD) is considered to be a powerful tool for generating time–frequency (TF) spectra and estimating the temporal and spectral features of this type of complex signal. However, due to its quadratic nature, undesirable cross-terms and spurious energies are also generated, which limit the readability and the interpretation of the spectrum. To suppress this effect, this paper describes the use of an autoregressive model based upon Burg’s maximum entropy method to modify the kernel of the discrete WVD. This technique was applied to ultrasonic Lamb wave signals obtained using various numerical and experimental configurations to extract the useful discriminatory spectral and temporal information required for the mode identification, localization, and quantification of structural damage. When used for damage localization, the proposed method precisely estimated the distance between two closely spaced notches in a metallic plate from different simulated noisy signals with a maximum uncertainty of 5%, based upon an excellent TF energy distribution. Moreover, the energy concentration (EC) of the TF energy distribution, in combination with the variation of its instantaneous frequency (IF) curve, was also effective in identifying the overlapping modes of the Lamb wave signal. Finally, for damage quantification, three TF-based damage indices, namely EC, TF flux, and IF, were extracted from five sets of specimens using the proposed TF scheme and used to train the regression model. Model testing proved that the damage indices have the potential to predict crack sizes precisely and reliably.

Synthetic environments for vision-based structural condition assessment of Japanese high-speed railway viaducts

Civil infrastructure condition assessment using visual recognition methods has shown significant potential for automating various aspects of the problem, including identification and localization of critical structural components, as well as detection and quantification of structural damage. The application of those methods typically requires large amounts of training data that consists of images and corresponding ground truth annotations. However, obtaining such datasets is challenging, because the images are annotated manually in most existing approaches. With the limited availability of datasets, development of effective visual recognition systems that can extract all required information is not straightforward. This research leverages synthetic environments to develop a unified system for automated vision-based structural condition assessment that can identify and localize critical structural components, and then detect and quantify damage of those components. The synthetic environments can produce images and associated ground truth annotations for semantic segmentation of structural components and damage, as well as monocular depth estimation for structural component localization. To illustrate the approach, automated vision-based structural condition assessment of reinforced concrete railway viaducts for a Japanese high-speed railway line (the Tokaido Shinkansen) is explored. The effectiveness of the synthetic environments and the generated dataset (the Tokaido dataset) is demonstrated by training fully convolutional network-based semantic segmentation and monocular depth estimation algorithms, and then testing the networks using both synthetic and real-world images. Finally, all trained algorithms are combined to realize an automated system for structural condition assessment.

Deep Learning-Based Concrete Surface Damage Monitoring Method Using Structured Lights and Depth Camera.

Due to the increase in aging structures and the decrease in construction workforce, there is an increasing interest in automating structural damage monitoring. Surface damage on concrete structures, such as cracks, delamination, and rebar exposure, is one of the important parameters that can be used to estimate the condition of the structure. In this paper, deep learning-based detection and quantification of structural damage using structured lights and a depth camera is proposed. The proposed monitoring system is composed of four lasers and a depth camera. The lasers are projected on the surface of the structures, and the camera captures images of the structures while measuring distance. By calculating an image homography, the captured images are calibrated when the structure and sensing system are not in parallel. The Faster RCNN (Region-based Convolutional Neural Network) with Inception Resnet v2 architecture is used to detect three types of surface damage: (i) cracks; (ii) delamination; and (iii) rebar exposure. The detected damage is quantified by calculating the positions of the projected laser beams with the measured distance. The experimental results show that structural damage was detected with an F1 score of 0.83 and a median value of the quantified relative error of less than 5%.

Integrated electromechanical impedance technique with convolutional neural network for concrete structural damage quantification under varied temperatures

The electromechanical impedance/admittance (inverse of impedance, EMI/EMA) technique has demonstrated its high sensitivity to structural incipient damages, however, it is also vulnerable to ambient temperature fluctuation that may adversely cause false alarms in real-life monitoring applications. This paper proposed an innovative approach integrating the EMA technique with Convolutional Neural Network (CNN) to quantify concrete structural damage severity under varied temperatures. In the approach, EMA signals were first split into multiple sub-range responses, and the corresponding statistical indices namely Pearson Correlation Coefficient (PCC), were calculated and utilized as input of one-dimensional CNN. In order to meet high accuracy for training the CNN model, an Orthogonal Matching Pursuit (OMP) algorithm based EMA data reconstruction technique was employed to generate sufficient data for the establishment of CNN data library. Effectiveness of the proposed connectionist approach was verified through crossover experiments of diagnosing crack damages on a standardized concrete cube subjected to varied temperatures. Both temperature recognition and damage quantification under elevated temperatures were accomplished via performing the EMA signal based CNN procedure. Experimental results revealed that the proposed approach was of high accuracy for temperature recognition and damage quantification, which potentially facilitates the in-situ monitoring of engineering structures under varied temperature conditions.

Damage Identification in Plate Structures Using Sparse Regularization Based Electromechanical Impedance Technique.

This paper proposes a novel structural damage quantification approach using a sparse regularization based electromechanical impedance (EMI) technique. Minor structural damage in plate structures by using the measurement of only a single surface bonded lead zirconate titanate piezoelectric (PZT) transducer was quantified. To overcome the limitations of using model-based EMI based methods in damage detection of complex or relatively large-scale structures, a three-dimensional finite element model for simulating the PZT–structure interaction is developed and calibrated with experimental results. Based on the sensitivities of the resonance frequency shifts of the impedance responses with respect to the physical parameters of plate structures, sparse regularization was applied to conduct the undetermined inverse identification of structural damage. The difference between the measured and analytically obtained impedance responses was calculated and used for identification. In this study, only a limited number of the resonance frequency shifts were obtained from the selected frequency range for damage identification of plate structures with numerous elements. The results demonstrate a better performance than those from the conventional Tikhonov regularization based methods in conducting inverse identification for damage quantification. Experimental studies on an aluminum plate were conducted to investigate the effectiveness and accuracy of the proposed approach. To test the robustness of the proposed approach, the identification results of a plate structure under varying temperature conditions are also presented.

Development and application of random forest technique for element level structural damage quantification

This paper presents element level structural damage quantification using an ensemble-based machine learning technique, namely, random forest technique, with acceleration responses from structures. The ensemble-based approach provides a better prediction than an individual model. Random forest is a machine learning algorithm which has several decision trees to perform a task. The proposed approach develops a random forest as a regressor to predict multiple output variables, which is the vector of elemental level damage quantification results of the structure. Damage severity is identified as the reduction in the elemental stiffness parameters. The acceleration responses for single-element and multiple element damage cases are generated and further processed to feed as input to the random forest. The acceleration responses from the sensor nodes are concatenated, and principal component analysis (PCA) is applied to reduce the uncorrelated input dimension. The proposed approach can provide good damage identification results efficiently with much less computational demand and time, compared to the neural network training methods and deep learning models. Moreover, fewer sensors are needed to measure acceleration responses to localise damage and quantify the damage level. To demonstrate the proposed method, a simply supported beam is used as an example in numerical studies. Different levels of noise in the acceleration responses and uncertainty in the finite element modelling are considered. Experimental studies on a steel frame structure with 70 elements are also conducted to investigate the performance of the proposed approach for structural damage quantification. Good identification results are obtained with an efficient training process.

Real-time structural health monitoring of nonlinear hysteretic structures using a fast and slow dynamics separation method

Structural health monitoring (SHM) is the process of identification, localisation, and quantification of structural damage due to external loads, such as an earthquake. SHM results simplify and automate typical visual or localised experimental approaches and enable more informed structural safety assessment and post-event retrofit. On-line SHM is of particular interest for rapid safety assessment by owners and civil defence authorities, and can also inform active control systems to further avoid or mitigate damage during an event. This significant advantage over off-line SHM has drawn many researchers to on-line or real-time methods. This paper presents an on-line SHM algorithm for model-based identification and monitoring of non-linear hysteretic structures. It separates fast and slow dynamics and uses robust piecewise least squares fitting over different time intervals to identify key structural parameters over time, including stiffness, damping, and the well-accepted Bouc–Wen hysteretic model parameters, which are directly related to well-recognised damage metrics. Proof-of-method simulations of a realistic nonlinear case study structure, subjected to a suite of 20 different ground motions, show the algorithm accurately identifies structural parameters to within 2.7% and 4.3% of the actual as-modelled values for stiffness and damping, respectively. Results for a range of major changes (damage) in structural parameters show the algorithm performs well in tracking these changes in real-time. Overall, the results present an accurate, real-time, model-based SHM method for highly nonlinear structures.