Structural Damage Detection Using Empirical Mode Decomposition: Experimental Investigation

This paper presents an experimental investigation on the applicability of the empirical mode decomposition (EMD) for identifying structural damage caused by a sudden change of structural stiffness. A three-story shear building model was constructed and installed on a shaking table with two springs horizontally connected to the first floor of the building to provide additional structural stiffness. Structural damage was simulated by suddenly releasing two pretensioned springs either simultaneously or successively. Various damage severities were produced using springs of different stiffness. A series of free vibration, random vibration, and earthquake simulation tests were performed on the building with sudden stiffness changes. Dynamic responses including floor accelerations and displacements, column strains, and spring releasing time instants were measured. The EMD was then applied to measured time histories to identify damage time instant and damage location for various test cases. The comparison of identified results with measured ones showed that damage time instants could be accurately detected in terms of damage spikes extracted directly from the measurement data by EMD. The damage location could be determined by the spatial distribution of the spikes along the building. The influence of damage severity, sampling frequency, and measured quantities on the performance of EMD for damage detection was also discussed.     

Nonlinear Identification of Lumped-Mass Buildings Using Empirical Mode Decomposition and Incomplete Measurement

Most structures exhibit some degrees of nonlinearity such as hysteretic behavior especially under damage. It is necessary to develop applicable methods that can be used to characterize these nonlinear behaviors in structures. In this paper, one such method based on the empirical mode decomposition (EMD) technique is proposed for identifying and quantifying nonlinearity in damaged structures using incomplete measurement. The method expresses nonlinear restoring forces in semireduced-order models in which a modal coordinate approach is used for the linear part while a physical coordinate representation is retained for the nonlinear part. The method allows the identification of parameters from nonlinear models through linear least-squares. It has been shown that the intrinsic mode functions (IMFs) obtained from the EMD of a response measured from a nonlinear structure are numerically close to its nonlinear modal responses. Hence, these IMFs can be used as modal coordinates as well as provide estimates for responses at unmeasured locations if the mode shapes of the structure are known. Two procedures are developed for identifying nonlinear damage in the form of nonhysteresis and hysteresis in a structure. A numerical study on a seven-story shear-beam building model with cubic stiffness and hysteretic nonlinearity and an experimental study on a three-story building model with frictional magnetoreological dampers are performed to illustrate the proposed method. Results show that the method can quite accurately identify the presence as well as the severity of different types of nonlinearity in the structure.     

Multimetric Sensing for Structural Damage Detection

Vibration-based damage detection methods have been widely studied for structural health monitoring of civil infrastructure. Acceleration measurements are frequently employed to extract the dynamic characteristics of the structure and locate damage because they can be obtained conveniently and possess relatively little noise. However, considering the fact that damage is a local phenomenon, the sole use of acceleration measurements that are intrinsically global structural responses limits damage detection capabilities. This paper investigates the possibility of using both global and local measurements to improve the accuracy and robustness of damage detection methods. A multimetric approach based on the damage locating vector method is proposed. Numerical simulations are conducted to verify the efficacy of the proposed approach.     

Identification of Parametric Variations of Structures Based on Least Squares Estimation and Adaptive Tracking Technique

An important objective of health monitoring systems for civil infrastructures is to identify the state of the structure and to detect the damage when it occurs. System identification and damage detection, based on measured vibration data, have received considerable attention recently. Frequently, the damage of a structure may be reflected by a change of some parameters in structural elements, such as a degradation of the stiffness. Hence it is important to develop data analysis techniques that are capable of detecting the parametric changes of structural elements during a severe event, such as the earthquake. In this paper, we propose a new adaptive tracking technique, based on the least-squares estimation approach, to identify the time-varying structural parameters. In particular, the new technique proposed is capable of tracking the abrupt changes of system parameters from which the event and the severity of the structural damage may be detected. The proposed technique is applied to linear structures, including the Phase I ASCE structural health monitoring benchmark building, and a nonlinear elastic structure to demonstrate its performance and advantages. Simulation results demonstrate that the proposed technique is capable of tracking the parametric change of structures due to damages.     

Structural Health Monitoring and Damage Assessment Using Frequency Response Correlation Criteria

Two frequency response correlation criteria, namely the global shape correlation (GSC) function and the global amplitude correlation (GAC) function, are established tools to quantify the correlation between predictions from a finite-element (FE) model and measured data for the purposes of FE model validation and updating. This paper extends the application of these two correlation criteria to structural health monitoring and damage detection. In addition, window-averaged versions of the GSC and GAC, namely WAIGSC and WAIGAC, are defined as effective damage indicators to quantify the change in structural response. An integrated method of structural health monitoring and damage assessment, based on the correlation functions and radial basis function neural networks, is proposed and the technique is applied to a bookshelf structure with 24 measured responses. The undamaged and damaged states, single and multiple damage locations, as well as damage levels, were successfully identified in all cases studied. The ability of the proposed method to cope with incomplete measurements is also discussed.     

Structural Damage Diagnosis and Assessment under Seismic Excitations

This study proposes a method of detecting, locating, and quantifying structural damage by directly using structural vibration measurements in the time domain. In this method, the coupling effect of the damage at different locations in the structure on the structural vibratory responses is eliminated by projecting these measured quantities onto some specific subspaces. As a result, the structural system, generally modeled with multiple degree of freedom, is decomposed into several independent single-degree-of-freedom (SDOF) systems, every one of which is only associated with the damage at one certain location or region. A monitor is designed as an observer to detect the structural damage related to each SDOF system. A decision-making scheme is developed to correlate the monitor’s output to the occurrence of the damage. The severity of the damage is estimated with a traditional system identification method in an iterative way. The analysis of the effects of measurement noise is also included. Numerical examples are presented to demonstrate the effectiveness of the proposed method.     

Optimum Sensor Placement for StructuralDamage Detection

A method of optimizing sensor locations and detecting damage in a structure using the collected information is presented in this paper. Unlike most methods found in the literature, the sensor locations are prioritized according to their ability to localize structural damage based on the eigenvector sensitivity method. This is in line with the requirements for structural damage localization. Only a small subset of the total structural degrees-of-freedom is instrumented, and the incomplete modes yielded from these optimized sensor locations are used to localize structural damage. Numerical example and test results show that this approach is effective for detecting structural damage directly using optimum and incomplete test modes.     

Formulation of Seismic Fragilities for a Wood-Frame Building Based on Visually Determined Damage Indexes

The study of earthquake engineering has made significant strides over the last one-half century with scientists developing methods to better understand the basis and mechanisms of earthquakes and engineers working to mitigate economic loss and fatalities. A paradigm known as performance-based seismic design (PBSD) not only provides life safety to building occupants, but seeks to control structural and nonstructural damage in buildings and other structures. The development of fragility curves based on the well-known Park-Ang damage index is examined herein. This type of formulation can provide the information needed to assess the seismic vulnerability of a structure. Existing shake table test data from the NEESWood Project’s test of a 223?m2 (1,800 sq ft) two-story house was combined with a participant survey to calibrate a damage model. The result was the development of damage fragilities based exclusively on nonlinear time history analysis. Then, the proposed numerical damage model was applied and fragility curves were developed for a six-story light-frame wood condominium building. The results appear logical based on observations of system-level shake table tests over the last decade, and thus the method shows promise provided significant torsion is not present in the system.     

Damage Detection based on Single-Input-Single-Output Measurements

A methodology is presented for detecting damage of structural systems maintaining a linear response. A single frequency response function measured at several frequencies along with a correlated analytical model of the undamaged structure are used to detect and assess damage. The method is directed toward situations where the number of damaged elements is generally known to be limited. A computationally efficient method of recalculating a single receptance is presented. Numerical results for a two-dimensional structural frame are presented to validate and assess the proposed approach. Issues for the development of the approach are discussed.     

Modeling of Temperature–Frequency Correlation Using Combined Principal Component Analysis and Support Vector Regression Technique

A good understanding of environmental effects on structural modal properties is essential for reliable performance of vibration-based damage diagnosis methods. In this paper, a method of combining principal component analysis (PCA) and support vector regression (SVR) technique is proposed for modeling temperature-caused variability of modal frequencies for structures instrumented with long-term monitoring systems. PCA is first applied to extract principal components from the measured temperatures for dimensionality reduction. The predominant feature vectors in conjunction with the measured modal frequencies are then fed into a support vector algorithm to formulate regression models that may take into account thermal inertia effect. The research is focused on proper selection of the hyperparameters to obtain SVR models with good generalization performance. A grid search method with cross validation and a heuristic method are utilized for determining the optimal values of SVR hyperparameters. The proposed method is compared with the method directly using measurement data to train SVR models and the multivariate linear regression (MLR) method through the use of long-term measurement data from a cable-stayed bridge. It is shown that PCA-compressed features make the training and validation of SVR models more efficient in both model accuracy and computational costs, and the formulated SVR model performs much better than the MLR model in generalization performance. When continuously measured data is available, the SVR model formulated taking into account thermal inertia effect can achieve more accurate prediction than that without considering thermal inertia effect.     

Natural Excitation Technique and Eigensystem Realization Algorithm for Phase I of the IASC-ASCE Benchmark Problem: Simulated Data

A benchmark study in structural health monitoring based on simulated structural response data was developed by the joint IASC–ASCE Task Group on Structural Health Monitoring. This benchmark study was created to facilitate a comparison of various methods employed for the health monitoring of structures. The focus of the problem is simulated acceleration response data from an analytical model of an existing physical structure. Noise in the sensors is simulated in the benchmark problem by adding a stationary, broadband signal to the responses. A structural health monitoring method for determining the location and severity of damage is developed and implemented herein. The method uses the natural excitation technique in conjunction with the eigensystem realization algorithm for identification of modal parameters, and a least squares optimization to estimate the stiffness parameters. Applying this method to both undamaged and damaged response data, a comparison of results gives indication of the location and extent of damage. This method is then applied using the structural response data generated with two different models, different excitations, and various damage patterns. The proposed method is shown to be effective for damage identification. Additionally the method is found to be relatively insensitive to the simulated sensor noise.     

Cross-Modal Strain Energy Method for Estimating Damage Severity

A newly developed damage severity estimation method, termed as cross-modal strain energy (CMSE) method, which is capable of accurately estimating the damage magnitude of multiple damaged members, is presented. While all existing damage severity estimation methods that utilize modal strain energy are either employing an iterative solution procedure or involving significant approximations, the CMSE method is an exact, noniterative solution method. Furthermore, the development of the CMSE method is under the assumption that the mass distributions of the baseline and damaged structures are unknown, but identical. Implementing this method requires only the information of a few modes measured from the damaged structure. Numerical studies are demonstrated for a three-dimensional five-story frame structure based on synthetic data generated from finite element models.     

Model-Based Damage Identification in a Continuous Bridge Using Vibration Data

Damage often causes changes in the dynamic characteristics of a structure such as frequencies and mode shapes. Vibration-based damage identification techniques utilize the changes in the dynamic characteristics of a structure to determine the location and extent of damage in the structure. Such techniques are applied in this study to the Crowchild Bridge, a steel-free deck continuous bridge located in western Canada. While the numerical models of the bridge are correlated with the measured dynamic characteristics, computer simulation is used to study the identification of a number of different damage patterns, and the effects of measurement errors and incomplete mode shapes on the quality of results are evaluated. The effectiveness of some selected damage identification techniques is examined; the potential difficulties in identifying the damage are outlined; and areas of further research are suggested. A three-dimensional finite-element model and a simple two-dimensional girder model of the bridge have been constructed to study the usefulness of the selected damage identification methods. Another promising damage detection method proposed here is based on the application of neural networks that combines a vibration-based method.     

Damage Detection in Composite Materials Using Identification Techniques

The present paper describes an approach for damage detection in composite structures that has its basis in methods of system identification. Response of a damaged structure differs from predictions obtained from a mathematical model of the original structure, where such a model is typically a finite‐element representation of the structure. In the present work dealing with composite materials, two distinct analytical models, one using two‐dimensional (2D) elements in conjunction with the classical lamination theory and another using three‐dimensional (3D) elements were considered. The output error approach of system identification was employed to determine changes in the analytical model necessary to minimize differences between the measured and predicted response. The proposed method is an extension of the stiffness‐reduction approach for damage detection to realistic structures. Numerical simulation of measurements of static deflections, strains, and vibration modes were used in the identification procedure. The methodology was implemented for representative composite structures. Principal shortcomings in the proposed approach and possible methods to circumvent these problems are discussed in the paper.     

Damage Identification Using Committee of Neural Networks

This paper presents a committee of neural networks technique, which employs frequency response functions (FRFs), modal properties (natural frequencies and mode shapes), and wavelet transform (WT) data simultaneously to identify damage in structures. The experimental demonstration of the method is obtained by studying the sensitivities of the FRFs, modal properties, and WT data to four types of faults in a cylindrical shell. The experimental results show that different faults affect data in a different manner. The proposed approach is tested on simulated data from a three-degree-of-freedom mass-spring-damper system. The results from the simulated study show that the performance of the approach is not influenced by the noise in the data. Finally, the method is used to identify damage in a population of ten steel seam-welded cylindrical shells. The proposed method is able to identify damage cases better than the three approaches used individually. The committee approach gives results that generally have a lower mean square error (MSE) than the average MSE of the individual methods. It is found that the effectiveness of the method is enhanced when experimentally measured data are used, which is in contrast to many existing methods. This is because the committee approach assumes that the errors given by the three approaches are uncorrelated, a situation that becomes more apparent when using measured data rather than simulated data.     

Structural Damage Detection and Assessment from Modal Response

This paper presents a global damage detection and assessment algorithm based on a parameter estimation method using a finite-element model and the measured modal response of a structure. Damage is characterized as a reduction of the member constitutive parameter from a known baseline value. An optimization scheme is proposed to localize damaged parts of the structure. The algorithm accounts for the possibility of multiple solutions to the parameter estimation problem that arises from using spatially sparse measurements. Errors in parameter estimates caused by sensitivity to measurement noise are reduced by selecting a near-optimal measurement set from the data at each stage of the localization algorithm. Damage probability functions are computed upon completion of the localization process for candidate elements. Monte Carlo methods are used to compute the required probabilities based on the statistical distributions of the parameters for the damaged and the associated baseline structure. The algorithm is tested in a numerical simulation environment using a planar bridge truss as a model problem.     

Statistical Damage Assessment of Framed Structures from Static Responses

This paper presents a damage assessment algorithm for framed structures based on a system identification scheme with a regularization technique. The regularization technique is introduced to alleviate the ill-posedness of the system identification problems. A new regularization function based on the Frobenius norm of the difference between the estimated and the baseline stiffness matrix is proposed. A parameter grouping technique is adopted to locate damaged members and to overcome sparseness of measured data. A data perturbation method is employed to obtain statistical distributions of system parameters with a set of noise-polluted measured data. A statistical approach by a hypothesis test is presented to assess damage. The validity of the proposed method is demonstrated through numerical examples.     

Application of a Statistical Model Updating Approach on Phase I of the IASC-ASCE Structural Health Monitoring Benchmark Study

This paper addresses the problem of structural health monitoring (SHM) and damage detection based on a statistical model updating methodology which utilizes the measured vibration responses of the structure without any knowledge of the input excitation. The emphasis in this paper is on the application of the proposed methodology in Phase I of the benchmark study set up by the IASC–ASCE Task Group on structural health monitoring. Details of this SHM benchmark study are available on the Task Group web site at 〈http://wusceel.cive.wustl.edu/asce.shm〉. The benchmark study focuses on important issues, such as: (1) measurement noise; (2) modeling error; (3) lack of input measurements; and (4) limited number of sensors. A statistical methodology for model updating is adopted in this paper to establish stiffness reductions due to damage. This methodology allows for an explicit treatment of the measurement noise, modeling error, and possible nonuniqueness issues characterizing this inverse problem. The paper briefly describes the methodology and reports on the results obtained in detecting damage in all six cases of Phase I of the benchmark study assuming unknown (ambient) data. The performance, limitations, and difficulties encountered by the proposed statistical methodology are discussed.     

Structural Damage Detection from Modal Strain Energy Change

A structural damage detection method based on modal strain energy (MSE) change before and after damage is presented in this paper. The localization of damage based on MSE of each structural element is briefly presented, and the sensitivity of the MSE with respect to a damage is derived. The sensitivity is not based on any series expansion and is a function of the analytical mode shape changes and the stiffness matrix. Only incomplete measured mode shapes and analytical system matrices are required in this damage localization and quantification approach. Results from a numerical example and an experiment on a single-bay, two-story portal steel frame structure are investigated. The effects of measurement noise and truncated analytical mode shapes are discussed. Results indicate that the proposed approach is noise sensitive, but it can localize single and multiple damages. Damage quantification of two damages is successful with a maximum of 14% error under a 5% measurement noise.     

Statistical methods for detecting activated regions in functional MRI of the brain

Two statistical tests for detecting activated pixels in functional MRI (fMRI) data are presented. The first test (t-test) is the optimal solution to the problem of detecting a known activation signal in Gaussian white noise. The results of this test are shown to be equivalent to the cross-correlation method that is widely used for activation detection in fMRI. The second test (F test) is the optimal solution when the measured data are modeled to consist of an unknown activation signal that lies in a known lower dimensional subspace of the measurement space with added Gaussian white noise. A model for the signal subspace based on a truncated trigonometric Fourier series is proposed for periodic activation-baseline imaging paradigms. The advantage of the second method is that it does not assume any information about the shape or delay of the activation signal, except that it is periodic with the same period as the activation-baseline pattern. The two models are applied to experimental echo-planar fMRI data sets and the results are compared.