Key Engineering Materials Vols. 413-414

Abstract: The aim of this work is the investigation and improvement of a Structural Health Monitoring method based on Lamb waves propagation. This research concentrates on ambiguity in damage localization using attached piezoelectric transducers as sources and sensors of the elastic waves. A linear phased array is chosen as a starting point of the investigation. It has a great advantage in damage localization, namely it enables to amplify the wave reflected from damage, increasing the signal to noise ratio, and precisely indicates not only the distance to damage from the array but also the direction on which the damage lies. However it has also a great disadvantage which needs to be handled – the localization results are symmetric in relation to the line on which the transducers of linear phased array are placed. This obviously does not facilitate Structural Health Monitoring process and precise indication of damage placement. Therefore this investigation aims to improve this localization method by removing the ambiguity in results. In this work the placement of transducers forming a linear phased array is modified to achieve this goal. Several array modification are investigated and compared in order to determine the best solution. Presented research is based on theoretical calculations as well as laboratory experiments on prepared specimens. The measurements are conducted with a compact 13–channel SHM system controlled by a MATLAB® script.

Abstract: In this paper algorithm for damage localisation in thin panels made of aluminium alloy has been proposed. Mentioned algorithm uses Lamb wave propagation methods and geometrical approach for damage localisation. Elastic waves are generated and received using piezoelectric transducers. Excited elastic waves propagate and reflect from panel boundary and discontinuities existing in the panel. Wave reflection can be registered through the piezoelectric transducers and used in signal processing algorithm. Processing algorithm consists of two parts: signal filtering and extraction of damage location. The first part is used in order to remove noise from received signals. Second part allows to extract arrival time of waves reflected from discontinuity, very often called Time Of Flight (TOF). Localisation algorithm uses pairs of transducers from a concentrated transducers configuration. Using signals from pair of transducers two times of reflection can be extracted from received signals. Because coordinates of transducers are well known ellipse can be constructed based on extracted times of waves reflections. Damage lies one ellipse but it is not known exactly where. Therefore one ellipse is not enough to localise a discontinuity. In order of proper damage localisation more ellipses must be used. In this purpose signals received by larger number of transducers pairs are used in damage localisation algorithm. Points of ellipses intersections allow to indicate localisation of damage. Described signal processing algorithm has been coded in the MATLAB® environment. In this work experimental results has been presented.

Abstract: According to the latest research results presented in the literature changes in propagating waves are one of the most promising parameters for damage identification algorithms. Numerous publications describe methods of damage identification based on the analysis of signals reflected from damage. They also include complicated signal processing techniques. Such methods work well for damage localisation, but it is rather difficult to use them in order to estimate the size of damage. It is natural that propagating wave reflects from any structural discontinuity. The bigger the disturbance the bigger part of a propagating wave reflects from it. The amount of energy reflected and transmitted through any discontinuity can expressed as reflection and transmission coefficients. In the literature different application for these coefficients may be found – the most often cited application is connected with localising changes in the geometry of structures. Changes in the coefficients due to cross section variations in rods and beams or due to existence of stiffeners in plates are well documented. However there are no application of using the reflection and transmission coefficients for damage size identification. For this reason the analysis presented in this paper has been carried out. The article presents a method of damage identification in 1D elements based on the wave propagation phenomenon and changes in reflection and transmission coefficients. The changes in transmission and reflection coefficients for waves propagating in isotropic rods with different types of damage have been analysed. The rods have been modelled with the elementary, two and three mode theories or rods. For numerical modelling the Spectral Finite Element Method has been used. Several examples are given in the paper.

Abstract: The aim of the work is to develop a procedure allowing the test engineer to determine the probability of finding a crack in a beam structure. The procedure is based on the use of wavelet analysis and the simulation is performed by taking advantage of spectral elements to represent accurately the dynamic behaviour of beam structures in the high frequency range. In this context, numerical analyses are performed with the final scope of simulating a real testing environment: measurement error is considered and spectral elements are used so as to avoid influencing the capacity of the procedure with regard to solving the inverse problem. In this article the relation between the excitation frequency and the probability of locating the fault is shown. In particular, it is demonstrated by simulation that the probability of correctly determining the fault location increases with the excitation frequency.

Abstract: Modernized cities must have adequate infrastructures to support the daily needs from her citizens. The continuity in pipeline services that supply water, gas and oil to citizens and deliver wastes to designated collectors are in prime concern to any modernized city. However, in-service pipeline is prone to defects due to aging, external impacts, or hazardous operating environment. It is of prime importance to apply an efficient inspection method to characterize the potential defect in pipeline so that the information of damage caused can be determined prior to the fatal rupture of pipeline. An early warning generated from an accurate characterization of defect can encourage the performance of proper remedy and maintenance for minimizing the scope of damage to pipelines. In this paper, a presentation is given to an advanced inspection technique based on ultrasonic guided waves. This technique has already shown great potentials in non-destructive testing of material and structures in many fields. The advantages and difficulties involved in the pipeline inspection using ultrasonic guided waves have been identified. For the quantitative characterization of defect in pipeline inspection based on advanced guided waves, we propose the method through considering the reflected signal since it provides useful information related to defect. The method analyzes the captured signals reflected from the potential defect, decomposes the embedded dimensional information of defect and then accordingly identifies its severity. Although the experiments were conducted on artificial defects, the results proved that qualitative characterization of defect is feasible. Combined with guided waves, our method can provide comprehensive information related to the existence, location, severity of defect etc., through the analysis of reflected signal from the interactions of excited guided waves with pipeline defect.

Abstract: The local effect of “softening” at the crack location can be simulated by an equivalent spring connecting the two segments of the beam. As modelling the crack, the non-perfectly rigid clamp is also simulated as a torsional spring of unknown stiffness. Combined with the Bernoulli-Euler theories of beam, the present model is applied to derive the characteristic equation of the cantilever beam under uncertain end conditions related to the crack parameters, namely, the location and the depth of the crack. Based on this characteristic equation, an accurate crack identification method is developed to identify the location and the depth of the crack by minimizing the difference between the analytical and experimental frequency values. The proposed approach is verified by two cantilever beam experiments under ideal boundary conditions and uncertain end conditions. It is found that the location and the depth of the crack can be worked out when at least three natural frequencies are known. For crack identification of the cantilever beam under uncertain end conditions, the identified crack location of the proposed approach is more accurate than the Narkis’ method. Furthermore, the crack depth can also be obtained by the present method.

Abstract: For detection of damage in frame and truss structures, the normalized cumulative energy is proposed as the identification parameter within the framework of the damage locating vector (DLV) method. Due to the limited number of sensors used, it is necessary to filter out the actual damaged elements from the identified set of potential damaged elements. An intersection scheme using only the measured signals is proposed for the filtering and verified using a warehouse frame comprising truss, beam and column elements. As wireless sensors are introduced into structural health monitoring systems, loss of data during transmission is one significant issue which needs to be addressed. An algorithm to patch the loss data is proposed and when integrated with the proposed damage detection method is experimentally shown to be feasible using a cantilever truss.

Abstract: Instantaneous impact signal analysis investigation and feature extraction have been broadly investigated by many researchers. How to determine an effective time-frequency distribution analysis method is urgent needed. In this paper, Hilbert spectrum frequency resolution is investigated to show signal typical features according to performance estimation model. Different frequency units for a Hilbert spectrum expression will be investigated on the effect of feature analysis in detail. An estimation model will be constructed to determine the best frequency resolution for signal analysis. By using this estimation, frequency resolution can be determined. This model can improve the performance of Hilbert spectrum to show signal features and accuracy for signal analysis and pattern recognition. To testify the effectiveness of this estimation model, numerical simulation will be used as an example to analyze the accuracy of this model. At the same time, different bearing working condition’s pattern recognition based on Hilbert spectrum will be used to testify the effectiveness of this model.

Abstract: Due to the fact that near a crack singularity, gradients of the solution are large and are also subject to abrupt changes, so that the solution cannot locally be accurately approximated by a piecewise polynomial function on a quasi-uniform mesh. Lifting wavelet finite element has good ability in modal analysis for singularity problems like a cracked pipe. The first three natural frequencies of the cracked pipe were solved with lifting wavelet finite element, and the database for crack diagnosis was obtained. The first three measured natural frequencies were employed as inputs and the intersection of the three frequencies contour lines predicted the normalized crack location and size. The experimental examples denote the method is of higher identification precision.

Abstract: The optimal selection of discriminatory features from large datasets remains a pressing problem in damage identification. In this paper, a Bayesian approach to classification and feature selection is introduced and applied to a challenging experimental problem.