Damage Detection Analysis Research Articles

Substructuring Technique for Damage Detection Using Statistical Multi-Stage Artificial Neural Network

Artificial Neural networks (ANN) have been proven in many studies to be able to efficiently detect damage from vibration measurements. Their capability to recognize patterns and to handle non-linear and non-unique problems provides an advantage over traditional mathematical methods in correlating the vibration data to damage location and severity. However, one shortcoming of ANN is they require enormous computational effort and sometimes prohibitive time and computer memory for training a reliable ANN model, especially when structures with many degrees of freedom are involved. Therefore, in most cases, rather large elements are used in the structure model to reduce the degrees of freedom. This results in the structural vibration properties not being sensitive to small damage in a large element. As a result, direct application of ANN to detecting damage in a large civil engineering structures is not feasible. In this study, a multi-stage ANN incorporating a probability method is proposed to tackle this problem. Through this method, a structure is divided into several substructures, and each substructure is assessed independently. In each subsequent stage, only the damaged substructures are analyzed, and eventually the location and severity of small structural damage can be detected. This approach greatly reduces the computational time and the required computer memory. Moreover, a probabilistic method is also used to include the uncertainties in vibration frequencies and mode shapes in damage detection analysis. It is found that this method reduces the uncertainty effect in frequencies due to duplication error in the multi-stage ANN model and reduces the uncertainty effect in mode shapes due to the damage in other substructures. The developed approach is applied to detect damage in numerically simulated and laboratory tested concrete slab. The results demonstrate that the proposed method can detect small damage with a higher level of confidence, and the undamaged elements are less likely to be falsely detected.

Modelling of Lamb wave interaction with open and closed fatigue cracks for damage detection

Lamb waves are the most widely used guided ultrasonic waves for structural damage detection. Lamb wave propagation in complex structures is very complicated due to multiple reflections and mode conversion at geometrical and material features. Numerical simulations can significantly ease wave propagation analysis for damage detection. The local interaction simulation approach is used for Lamb wave interaction with fatigues cracks in an aluminium plate. The results, investigated for various crack lengths, are validated experimentally. The study shows that Lamb wave amplitude and arrival time are different for fully open and closed fatigue cracks. As a result damage detection sensitivity could be affected.

Numerical simulations for impact damage detection in composites using vibrothermography

Composite materials are widely used in many engineering applications due to their high strength-to-weight ratios. However, it is well known that composites are susceptible to impact damage. Detection of impact damage is an important issue in maintenance of composite structures. Various non-destructive image-based techniques have been developed for damage detection in composite materials. These include vibrothermography that detects surface temperature changes due to heating associated with frictional energy dissipation by damage. In the present paper numerical simulations are used to investigate heat generation in a composite plate with impact damage in order to support damage detection analysis with vibrothermography. Explicit finite elements are used to model ultrasonic wave propagation in the damaged plate. Simulated delamination and cracks induce frictional heating in the plate. Coupled thermo-mechanical simulations are performed in high frequencies using commercial LS-Dyna finite element code. Very good qualitative agreement between measurements and simulations has been obtained. The area of increased temperature corresponds very well with the damaged area in both experiments and simulations. Numerical model has to be further refined in order to quantitatively match the experiments. The main issues of concern are frictional and thermal properties of composites. The final goal of these research efforts is to predict damage detection sensitivity of vibrothermography in real engineering applications based on numerical models.

구조안전성 평가를 위한 무선 진동 모니터링 시스템 개발

Wired monitoring systems have been used for damage detection and dynamic analysis of large structures(bridges, dams, plants, etc.). However, the real-world applications still remain limited, mainly due to time and cost issues inherent to wired systems. In recent years, an increasing number of researchers have adopted WSN(wireless sensor network) technologies to the field of SHM(structural health monitoring). Accurate time synchronization is most critical for the wireless approach to be feasible for SHM purpose, along with sufficient wireless bandwidth and highly precise measuring resolution. To satisfy technical criteria stated above, a wireless vibration monitoring system that uses high-precision MEMS(micro-electro-mechanical system) sensors and A/D convertor is discussed in detail. It was found experimentally that the level of time synchronization fell within <TEX>$200\;{\mu}sec$</TEX>.

A Comparative Study of Statistical Outlier Analysis and ANN Methods for Damage Detection and Assessment in Composite Structures

This article introduces some of the experimental and analytical work in damage detection applied to polymeric composite T-joints used in maritime structures. Two methods of damage detection are discussed — A statistics-based outlier technique and one using artificial neural networks (ANNs). The SHM using ANNs system was found to be capable of not only detecting the presence of multiple delaminations in a composite structure, but also capable of determining the location and extent of all the delaminations present in the T-joint structure, regardless of the load (angle and magnitude) acting on the structure. The system developed, relies on the examination of the strain distribution of the structure under operational loading. Finally, on testing the SHM system developed with strain signatures of composite T-joint structures, subjected to variable loading, embedded with all possible damage configurations (including multiple damage scenarios), an overall damage (location and extent) prediction accuracy of 88.32% was achieved. These results are presented and discussed in detail in this article.

Dispersion analysis of Lamb waves and damage detection for aluminum structures using ridge in the time-scale domain

In this paper, the dispersion of Lamb waves in aluminum structures was systematically analyzed to differentiate the mode of each package in Lamb wave signals and localize damage. Piezoelectric transducers were bonded on the surfaces of aluminum structures, functioning as actuator and sensor to excite and acquire Lamb waves, respectively. Wavelet transform was applied to the acquired Lamb wave signals, in which the optimal mother wavelet was selected using the concept of Shannon entropy to obtain the most accurate location of each wave package. The ridge and contour of the Lamb wave signals in the time-scale domain were obtained to distinguish the mode of each wave package and pinpoint these packages for estimating the actual group velocities of dispersion curves and localizing damage. The proposed approach could help search the actual dispersion curves in the excitation frequency band by using only one Lamb wave signal. Ridges in the time-scale domain and the actual group velocities were further used to identify damage in the structures. Results demonstrate that the proposed approaches were effective in dispersion analysis, wave mode differentiation and damage localization.

Terrestrial Laser Scanning-Based Structural Damage Assessment

Terrestrial laser scanning (TLS) provides a rapid, remote sensing technique to model 3D objects. Previous work applying TLS to structural analysis has demonstrated its effectiveness in capturing simple beam deflections and modeling existing structures. This paper extends TLS to the application of damage detection and volumetric change analysis for a full-scale structural test specimen. Importantly, it provides a framework necessary for such applications, in combination with an analysis approach that does not require tedious development of complex surfaces. Intuitive slicing analysis methods are presented, which can be automated for rapid generation of results. In comparison with conventional photographic and surface analysis methods, the proposed approach proved consistent. Furthermore, the TLS data provided additional insight into geometric change not apparent using conventional methods. As with any digital record, a key benefit to the proposed approach is the resulting virtual test specimen, which is available for posttest analysis long after the original specimen is demolished. Uncertainties that can be introduced from large TLS data sets, mixed pixels and parallax in the TLS analysis are also discussed.

An outlier analysis of MFC-based impedance sensing data for wireless structural health monitoring of railroad tracks

This paper presents an outlier analysis for damage detection of railroad tracks using a macro-fiber composite (MFC) impedance-based wireless structural health monitoring (SHM) system. The impedance-based SHM method has some limitations because the measured impedance data may have considerable deviations caused by environmental or operational condition changes, including temperature, humidity, external loadings, or MFC patch bonding conditions. Thus, the method sometimes gives false-positive indication even for healthy structures. In order to overcome this limitation, an outlier analysis based on Mahalanobis squared distance (MSD) was proposed by taking root mean square deviation (RMSD) values of impedance signatures as a damage-sensitive feature vector. Optimal threshold values for both RMSD and MSD were determined through the proposed outlier analysis. The results showed that the use of MSD improved the damage detection capability with a lower threshold level as compared to that of RMSD. In this study, the applicability of the proposed method was experimentally verified by detecting three types of the railroad track damage, including head damage, web damage, and flange damage, which were simulated under the laboratory setting.

Hyperspectral imaging combined with principal component analysis for bruise damage detection on white mushrooms (Agaricus bisporus)

AbstractHyperspectral imaging (HSI) combines conventional imaging and spectroscopy to simultaneously acquire both spatial and spectral information from an object. This technology has recently emerged as a powerful process analytical tool for rapid, non‐contact and non‐destructive food analysis. In this study, the potential application of HSI for damage detection on the caps of white mushrooms (Agaricus bisporus) was investigated. Mushrooms were damaged by controlled vibration to simulate damage caused by transportation. Hyperspectral images were obtained using a pushbroom line‐scanning HSI instrument, operating in the wavelength range of 400–1000 nm with spectroscopic resolution of 5 nm. The effective resolution of the CCD detector was 580 × 580 pixels by 12 bits. Two data reduction methods were investigated: in the first, principal component analysis (PCA) was applied to the hypercube of each sample, and the second PC (PC 2) scores image was used for identification of bruise‐damaged regions on the mushroom surface; in the second method PCA was applied to a dataset comprising of average spectra from regions normal and bruise‐damaged tissue. In this case it was observed that normal and bruised tissue were separable along the resultant first principal component (PC 1) axis. Multiplying the PC 1 eigenvector by the hypercube data allowed reduction of the hypercube to a 2‐D image, which showed maximal contrast between normal and bruise‐damaged tissue. The second method performed better than the first when applied to a set of independent mushroom samples. The results from this study could be used for the development of a non‐destructive monitoring system for rapid detection of damaged mushrooms on the processing line. Copyright © 2008 John Wiley &amp; Sons, Ltd.

Noise analysis for sensitivity-based structural damage detection

As vibration-based structural damage detection methods are easily affected by environmental noise, a new statistic-based noise analysis method is proposed together with the Monte Carlo technique to investigate the influence of experimental noise of modal data on sensitivity-based damage detection methods. Different from the commonly used random perturbation technique, the proposed technique is deduced directly by Moore-Penrose generalized inverse of the sensitivity matrix, which does not only make the analysis process more efficient but also can analyze the influence of noise on both frequencies and mode shapes for three commonly used sensitivity-based damage detection methods in a similar way. A one-story portal frame is adopted to evaluate the efficiency of the proposed noise analysis technique.

Optimization of Wavelet Packets Analysis for Damage Detection in Composite Materials

A method based on entropy-based criteria is present to choose the optimal decomposition of Wavelet Packets Analysis (WPA) for damage detection in composite materials. The structural damage indexes constructed based on energy spectrum variation of the structural vibration responses decomposed using WPA before and after the occurrence of structural damage usually generate a complete binary tree to calculate its elements. Date mining is carried out in this paper by adoption entropy as the criteria to choose the optimal decomposition tree. In the decomposition process, only the sub-signals which contain main information of the original signal are decomposed to generate next level sub-signals. New damage index is constructed based on the optimal decomposition. Then the dimension of the damage index is reduced while still keeping its sensitive to damage. Whether Artificial Neural Network (ANN) or genetic algorithm (GA) is used in the further process of telling structural damage status from damage index, this reduction will make remarkable time saving.

Time–frequency and time–scale analyses for structural health monitoring

Signal processing is one of the most important elements of structural health monitoring. This paper documents applications of time-variant analysis for damage detection. Two main approaches, the time-frequency and the time-scale analyses are discussed. The discussion is illustrated by application examples relevant to damage detection.

Seismic Response and Damage Detection Analyses of an Instrumented Steel Moment-Framed Building

The seismic performance of steel moment-framed buildings has been of particular interest since brittle fractures were discovered at the beam–column connections in a number of buildings following the M 6.7 Northridge earthquake of January 17, 1994. A case study of the seismic behavior of an extensively instrumented 13-story steel moment frame building located in the greater Los Angeles area of California is described herein. Response studies using frequency domain, joint time–frequency, system identification, and simple damage detection analyses are performed using an extensive strong motion dataset dating from 1971 to the present, supported by engineering drawings and results of postearthquake inspections. These studies show that the building’s response is more complex than would be expected from its highly symmetrical geometry. The response is characterized by low damping in the fundamental mode, larger accelerations in the middle and lower stories than at the roof and base, extended periods of vibration after the cessation of strong input shaking, beating in the response, elliptical particle motion, and significant torsion during strong shaking at the top of the concrete piers which extend from the basement to the second floor. The analyses conducted indicate that the response of the structure was elastic in all recorded earthquakes to date, including Northridge. Also, several simple damage detection methods employed did not indicate any structural damage or connection fractures. The combination of a large, real structure and low instrumentation density precluded the application of many recently proposed advanced damage detection methods in this case study. Overall, however, the findings of this study are consistent with the limited code-compliant postearthquake intrusive inspections conducted after the Northridge earthquake, which found no connection fractures or other structural damage.

Damage Detection of Composite Structures Using Dynamic Analysis

A study on the use of modal parameter analysis for damage detection of structures made of composites is conducted. The damage-induced variations of modal parameters are investigated both numerically and experimentally. An appropriate finite element model is proposed to analyze the dynamic characteristics of different types of structures made of composites, such as honeycomb sandwich plates and multi-layer composite plates, with internal cracks and delamination. The numerical results are in good agreement with experimental results available in the literature. Natural frequencies, modal displacements, strains and energy are analyzed for the determination of damage severity and location. Vibration measurements are carried out using piezoelectric patch actuators and sensors for comparison and verification of the FEM model proposed in this study. Energy spectrum for wavelet packets decomposition of structural dynamic responses is used to highlight the features of damaged samples. The mechanism of mode-dependent energy dissipation of composite plates due to delamination is revealed for the first time. Experimental results clearly show the dependence of changes of modal parameters on damage size and location. The results obtained in this study show that the measured modal damping change combined with the computed modal strain energy distribution can be used to determine the location of delamination in composite structures. Both numerical and experimental findings in this study are significant to the establishment of guideline for size and location identification of damage in composite structures.

A probabilistic approach to structural damage estimation

A framework for probabilistic estimation of structural damage, which includes identifying its existence, location, and severity, through the system identification of a building, utilizing its observed vibration data in undamaged and damaged states, is presented in this paper. The method is applied to the vibration data obtained by the shaking table tests of a five-storey steel frame experimental model, to which many kinds of damaged states are given. The results from the damage detection analysis show the following: The method estimates the storey stiffness of the frame model very accurately. Even the damaged storeys corresponding to the stiffness reduction ratio of 1 or 2% are identified. The reliability of the estimated values are evaluated quantitatively and are reflected in the damage probability. This is very effective for the application to the health monitoring which requires the appropriate judgment on the damage states. Copyright © 2005 John Wiley & Sons, Ltd.

Application of Wavelet Approach for ASCE Structural Health Monitoring Benchmark Studies

This paper presents an application of wavelet analysis for damage detection and locating damage region(s) for the ASCE structural health monitoring benchmark data. The response simulation data were generated basically by a FEM program provided by the ASCE Task Group on Health Monitoring for a four-story prototype building structure subjected to simulated stochastic wind loading. Damage was introduced in the middle of response by breaking one or more structure elements such as interstory braces. Wavelets were used to analyze the simulation data. It was found that structural damage due to sudden breakage of structural elements and the time when it occurred can be clearly detected by spikes in the wavelet details. The damaged region can be determined by the spatial distribution pattern of the observed spikes. The effects of measurement noise and the severity of damage were investigated. The results in this paper illustrate a great promise of wavelet analysis for structural health monitoring, especially for an on-line application.

Damage Assessment of a Nonlinear Space Antenna Structure Using Quasi-Linearization Methods

Health monitoring of structural systems for the detection of damage is a critical practice in determining their safety and service life. A large amount of research has been done in applying damage detection analysis to linear models of structures. Although many real structures exhibit nonlinear behavior before or as a result of damage, there are currently few tools to apply damage detection analysis to structures that are modeled by nonlinear differential equations. The objective of this study is to initiate a facet of damage detection analysis applied to such nonlinear models of structures. Given a parameterized ordinary differential equation model of a structural system (linear or nonlinear), parameter estimation can serve as an assessment of possible or known damage to the structure. A numerical analysis of estimating parameters in a nonlinear structural model is investigated by applying a quasi-linearization method. The analysis is successfully performed on a nonlinear eight-degree-of-freedom space antenna model.

フィルタリングアルゴリズムを用いた大型浮遊式海洋建築物模型の損傷同定解析

This paper is concerned with the identification analysis of structural damage detection on the unit-linked floating structure model based on fluid-coupled free vibration equation and filtering algorithm. The problem of structural damage detection is formulated as an inverse problem in structural engineering. The formulated inverse problem must be analyzed under the consideration of stochastic properties on the responses of system because the observations of the system are usually measured in the presence of noise. The filtering algorithm has been well known as the solution method that is able to consider stochastic properties. In this paper two kinds of filtering algorithm, namely the extended Kalman filtering algorithm and new filtering algorithm based on the projection filter, are employed in the identification analysis. The coupled natural frequencies used as the observations and stochastic properties of system responses adopted the data obtained by experimental modal analysis.

Detecting Jack Pine Budworm Defoliation Using Spectral Mixture Analysis: Separating Effects from Determinants

Insect defoliation is a major disturbance force in forested ecosystems. Monitoring outbreaks, and estimating the areas affected, is therefore important for both forest managers and forest ecologists. The objective of our study was to classify jack pine budworm defoliation levels in Landsat TM imagery recorded previous to and during the 1990–1995 outbreak in our 450,000 ha study area in northwestern Wisconsin (USA). Many previous studies correlated insect defoliation and remotely sensed imagery with moderate to good success, but it often remained unclear if actual defoliation effects or other forest attributes correlated with defoliation were detected. For example, a larger deciduous component in mixed jack pine stands will limit budworm populations and thereby defoliation. The deciduous component in stands is a determining factor for insect defoliation, whereas needle discoloration and tree mortality are their effect. We used pre-outbreak Landsat TM data (1987) to identify determining factors for jack pine budworm population levels and peak-outbreak imagery (1993) for detecting actual defoliation. Our satellite data were atmospherically corrected using a radiative transfer model (5S). Spectral mixture analysis was performed using spectrometer measurements of jack pine needles and bark as representations of surface materials (“endmembers”). The explanatory power of the resulting fraction images was evaluated using jack pine budworm population data collected at 33 sampling points. Near-infrared reflectance (NIR) increased in defoliated stands between 1987 and 1993, but single date NIR in each year was negatively correlated with budworm levels in 1993 ( r=−0.69 and −0.47). This was because hardwood trees within jack pine stands caused higher NIR reflectance but limited jack pine budworm populations. The 10% NIR difference between pure and mixed jack pine stands outweighed the 3–5% increase in NIR due to defoliation and necessitated stratification of the satellite data by tree species. Spectral mixture analysis performed on pure jack pine stands resulted in a strong negative correlation between the 1993 green needle fraction and the 1993 budworm population data ( r=−0.94). This study was, to our knowledge, the first that applied spectral mixture analysis for forest damage detection, and also the first to use insect population measurements as independent field data. These methods, and the separation of determinants and effects of jack pine budworm defoliation, enabled us to detect actual defoliation with high accuracy.

Structural and Mechanical Damage Detection Using Wavelets

Damage detection has become an essential part of various research and maintenance activities in mechanical, aerospace, and civil engineering. Many of the last advances in mechanical and structural damage detection are related to new developments in the area of advanced signal processing. Over the past 10 years, wavelet theory in particular has been one of the emerging and fast-evolving mathematical and signal processing tools for vibration analysis. This paper is an attempt to collate in one place some of the recent advances in wavelet analysis for damage detection.