No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Acoustic emission characterization of failure modes in composites with ANN
Abstract
Noise suppression in acoustic emission data was attempted by developing and using artificial neural networks (ANN) and with the long-term objective of in-flight monitoring. In-flight experiments conducted earlier and the noise characteristics outlined therein were taken as basis for their simulation in the laboratory. Simulated noise sources were classified through both supervised and a combination of un-supervised and supervised training of ANN. AE signals were generated by fatigue spectrum load tests on CFRP specimens and their failure modes were characterized. Finally, simulated noise and the actual signals were mixed and re-classified into their respective classes. The results obtained are encouraging and the methods and procedures adopted confirm the feasibility of the approach for field applications.
... These AE signals are often mingled with electric signals and artificial noises. Therefore, how to distinguish these signals becomes a significant topic in AE investigations (Yang et al., 2002;Yi et al., 2002;Bhat et al., 2003). ...
... The major benefits in using ANN are the excellent management of uncertainties, noisy data, and nonlinear relationships. Neural network modeling has become increasingly accepted and is an interesting method for application to the AE technique (Grabec and Kuljani c, 1994;Kwak and Song, 2001;Yi et al., 2002;Bhat et al., 2003;Kwak and Ha, 2004;Leone et al., 2006). ...
... Analysing the AE data obtained from fatigue tests on such composite materials therefore requires considerable attention to detail, and often involves a significant number of material and test parameters as well as consideration of both the nature and quantity of the data. Most studies investigating damage mechanisms in composite materials have used pattern recognition as a multivariable technique for AE event classification (Bar, Bhat, and Murthy 2004;Bhat, Bhat and Murthy 2003;Godin et al. 2004;Huguet et al. 2002;Philippidis, Nikolaidis, and Anastassopoulos 1998;Philippidis, Nikolaidis, and Kolaxis, 1999). Bar, Bhat, and Murthy (2004) conducted research using the AE technique to analyse the mechanisms by which damage arose in multi-layered glass fiber reinforced plastic. ...
... The AE signals were captured through a polyvinylidene fluoride film sensor as of composite laminates of three dissimilar sets stacking sequences during monotonically increasing tensile load. Bhat, Bhat and Murthy (2003) meanwhile used artificial neural networks to identify noise suppression in AE data, with the long-term objective of in-flight monitoring on airplanes. In contrast, very few writers have studied damage modes in the usage of AE distributions in glass fiber reinforced composite. ...
This study assesses the progression of damage occurring on glass fiber reinforced polyester composite specimens using acoustic emission (AE) parameters. Its aims are to improve understanding of the particular characteristics of AE signals; and also to determine the relationship between AE signals and the failure of the material. Time and frequency domain trends were analysed at four different applied loads (60.97, 67.75, 74.52 and 81.30 MPa) representing 45–60% of the ultimate tensile strength of material. The relevant AE parameters were analysed both in the early stages of the test and as the material neared the fracture zone. The results showed a high degree of correlation between the root mean square and number of hits AE values and the number of cycles to failure, of 92.99 and 92.19%, respectively. This correlation, as well as AE basic parameters, suggests that AE can be a valuable tool to predict the fatigue life and detect the onset of damage in such composite materials.
... Most studies that investigated about damage mechanism in composite materials used pattern recognition as a multivariable technique for AE event classification [3][4][5][6][7][8]. Bar,Bhat [8] conducted a research with the purpose of usage of the acoustic emission technique (AE) to classify the mechanism of damage in multi-layered glass fibers reinforced plastic (GFRP). ...
... The signals of AE were developed through a polyvinylidene fluoride (PVDF) film sensor as of composite laminates of three dissimilar sets stacking sequences during monotonically increasing tensile load. Bhat et al., [7] used artificial neural networks (ANN) for finding noise suppression in acoustic emission data and with the long-term objective of inflight monitoring. ...
This paper presents acoustic emission (AE) technique for detecting onset damage of composite materials damage and validate this technique using actual AE data from fatigue crack growth. AE piezoelectric transducer was attached to glass fibre reinforced polyester composite specimen during the fatigue cyclic test. For data collection, AE parameters, i.e., duration, amplitude, and energy near fracture zone were obtained and were correlated to fatigue life. AE signals were obtained at four different applied loads (60.97MPa, 67.75MPa, 74.52MPa, and 81.30 MPa) which were 45% to 60% of ultimate tensile strength (UTS) of material. The results show correlation between AE parameters and the number of cycles to failure. This correlations show that AE can be used to predict the fatigue life and can be tool for detecting damages in composite materials.
... These AE signals are often mingled with electric signals and artificial noises. Therefore, how to distinguish these signals becomes a significant topic in AE investigations [3][4][5]. ...
... The major benefits in using ANN are the excellent management of uncertainties, noisy data, and nonlinear relationships. Neural network modeling has become increasingly accepted and is an interesting method for application to the AE technique [4,5,[12][13][14][15]. ...
Different types of rocks generate acoustic emission (AE) signals with various frequencies and amplitudes. How to determine rock types by their AE characteristics in field monitoring is also useful to understand their mechanical behaviors. Different types of rock specimens (granulite, granite, limestone, and siltstone) were subjected to uniaxial compression until failure, and their AE signals were recorded during their fracturing process. The wavelet transform was used to decompose the AE signals, and the artificial neural network (ANN) was established to recognize the rock types and noise (artificial knock noise and electrical noise). The results show that different rocks had different rupture features and AE characteristics. The wavelet transform provided a powerful method to acquire the basic characteristics of the rock AE and the environmental noises, such as the energy spectrum and the peak frequency, and the ANN was proved to be a good method to recognize AE signals from different types of rocks and the environmental noises.
... Environment temperature and temperature rise are both reported to significantly affect the fatigue life of the composites under cyclic loading. [1][2][3][4][5][6] Therefore, non-destructive methods for in-situ monitoring of their integrity via AE sensors [7][8][9][10][11] and infrared (IR) thermography [12][13][14][15][16] are employed to study the damage evolution of the laminates. In addition to non-intrusive methods, there are fatigue damage approaches to model the degradation of the materials. ...
... (a) Cantilever composite beam; (b) one-to-one correlation for whole fatigue life; and (c) stiffness degradation using equation(9). ...
The effect of the surface cooling on the fatigue life of a glass/epoxy laminate during fully reversed bending tests is investigated both experimentally and analytically. The experimental tests involve the use of acoustic emission and infrared thermography to monitor the structural integrity and evolution of damage. An analytical study is also performed to calculate the stress in the outermost layer of the laminate for both cases of cooled and uncooled specimens. The results show that the life of the laminate is highly dependent on the temperature and that surface cooling, if done appropriately, can significantly increase the fatigue life of the laminate.
... While vibration-based techniques are the most commonly employed [93], acoustic emission and electro-mechanical impedance (EMI) are examples of sensing techniques that are often discussed in the context of damage detection within SHM for aerospace structures. The former uses sensors to detect high-frequency stress waves that are generated during crack propagation and, while aerospace applications have been considered since the 1970s [126], significantly enhanced accuracy of crack-location prediction has been achieved for aerospace applications using ML methods such as Gaussian process regression [131] and NNs [25] in recent years. Further examples of the use of ML methods being used for damage detection classification within the aerospace industry include probabilistic neural networks [115], support-vector machines [194] and logistic regression [146]. ...
This review covers the new developments in machine learning (ML) that are impacting the multi-disciplinary area of aerospace engineering, including fundamental fluid dynamics (experimental and numerical), aerodynamics, acoustics, combustion and structural health monitoring. We review the state of the art, gathering the advantages and challenges of ML methods across different aerospace disciplines and provide our view on future opportunities. The basic concepts and the most relevant strategies for ML are presented together with the most relevant applications in aerospace engineering, revealing that ML is improving aircraft performance and that these techniques will have a large impact in the near future.
... This research aims at the latter. Aircraft structures comprise a large number of bolts, fasteners and plates, which move relative to one another -as very well explained Bhat et al. (2003). This leads to bolt hole rubbing (friction noise) as well as crack face rubbing and fretting. ...
p>Acoustic Emission (AE0) is a well established Non-Destructive Testing and Evaluation technique for damage monitoring and flaw location and has been used in a wide variety of fields, such as the aerospace and nuclear industry. The material studied here is fibre reinforced composites, which are not monolithic and therefore can fail in different modes. The research determines whether a characterisation of the damage is possible in terms of AE. This is accomplished by studying classic AE features (such as duration, counts, etc) and the frequency content.
The experimental work described here, by using pencil lead breaks, assesses the suitability of the AE parameters to characterise a source. The work shows that a characterisation must deal with the effects of the material and lay-up, shows the effects that the dimensions of the sample have on the internal reflections of the elastic waves and ultimately on the recorded signals, analyses how the accumulation of the signal with the time can provide useful information, illustrates a compact way to present the typically large number of AE data coming from the testing of composites and shows that the signals coming from a single sensor can carry information on the geometry of the structure. Studies on tensile tests in CFRP strips of different lay-ups and one panel loaded with a four-point bending are also included, to test the performance and the feasibility of AE in charactering actual sources of damage.
The novelty of this work consists of the following points: the difference between a description and a characterisation was defined; it was shown that the characteristics of the sensors largely affect the description; the variability introduced by the system and the testing parameters were investigated; the importance of the non-stationarity of the signals was illustrated, together with how this can yield to new information.
It was concluded that a characterisation can only have a weak meaning.</p
... The application of machine learning in damage monitoring based on AE mainly focuses on the classification of health and damage status of composite structures, and most of the methods used in literatures are unsupervised learning, including K-means [128][129][130][131], PCA [132], GMM [133], SOM [134,135], C-means [136], etc. There are also some studies using supervised learning methods, such as KNN [137], ANN [138,139], SVM [140], Bayesian method [141], etc. Few studies have used semi-supervised learning methods (such as the density peak algorithm) to identify cracks in composite matrices [142]. ...
Composite materials have been widely used in many industries due to their excellent mechanical properties. It is difficult to analyze the integrity and durability of composite structures because of their own characteristics and the complexity of load and environments. Structural health monitoring (SHM) based on built-in sensor networks has been widely evaluated as a method to improve the safety and reliability of composite structures and reduce the operational cost. With the rapid development of machine learning, a large number of machine learning algorithms have been applied in many disciplines, and also are being applied in the field of SHM to avoid the limitations resulting from the need of physical models. In this paper, the damage monitoring technologies often used for composite structures are briefly outlined, and the applications of machine learning in damage monitoring of composite structures are concisely reviewed. Then, challenges and solutions for quantitative damage monitoring of composite structures based on machine learning are discussed, focusing on the complete acquisition of monitoring data, deep analysis of the correlation between sensor signal eigenvalues and composite structure states, and quantitative intelligent identification of composite delamination damage. Finally, the development trend of machine learning-based SHM for composite structures is discussed. © 2022 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
... The proposed method has successfully discriminated against a variety of failures, regardless of the variance of the AE signal parameters associated with such methods. ANN was used to eliminate noise of several types from AE signals and to set AE events which are three different ways to fail in CFRP types by Bhat et al.[23]. The AE single attributes used as a descriptor were the time of ascent, calculation, strength, and height. ...
There is a continuous quest in the research community for superior and more accurate methodology for fault diagnosis and condition monitoring of diverse composite structure. This is because, these structures suffer from various nonlinear mode of failures while in service those are recognised as delamination, voids, matrix crack etc. Early detection of failures is what the most research mainly aims at. In this regard, the implementation of Artificial Intelligence (AI) techniques has been proved to be a versatile method for damage assessment. The collective inevitable use of composite materials in various high-performance engineering industries requires preliminary testing (detection, location, and quantification) for damage to these materials in order to improve their integrity and order. The present paper aims to bring out a concise review on various methodologies employed for damage/fault detection in composite materials with a special emphasis on supervised and unsupervised machine learning techniques. The major observations are outlined with an objective to put forward a broad perspective of the state of art related to laminated composite structural heath monitoring.
... V.Kostopoulos, et al. [13,14] combined a variety of different algorithms and compared the results of various algorithms in order to obtain the optimal classification of damage signals of ceramic matrix composites. In general, the unsupervised pattern recognition [15][16][17][18][19] was based on AE activity of each class combined with existing understanding of the damage mechanism of composite materials to determine the relationship between various types of signals and damage mechanisms. V.Kostopoulos, et al. [13,14] established the relationship between damage modes and AE signals by analyzing different classes of signals and combining scanning electron microscope. ...
In this paper, multiscale acoustic emission (AE) signal analysis was applied to acoustic emission data processing to classify the AE signals produced during the tensile process of C/SiC mini-composites. An established unsupervised clustering algorithm was provided to classify an unknown set of AE data into reasonable classes. In order to correctly match the obtained classes of the AE signals with the damage mode of the sample, three scales of materials were involved. Single fiber tensile test and fiber bundle tensile test were firstly performed to achieve the characteristics of AE signal of fiber fracture. Parameter analysis and waveform analysis were added to extract the different features of each class of signals in the In-situ tensile test of C/SiC mini-composite. The change of strain field on the sample surface analyzed by DIC (Digital Image Correlation) revealed the corresponding relationship between matrix cracking and AE signals. Microscopic examinationwas used to correlate the clusters to the damage mode. By analyzing the evolution process of signal activation for each class against the load, it also provided a reliable basis for the correlation between the obtained classes of the AE signals and the damage mechanism of the material.
... A large number of papers have demonstrated that AE technology, as a passive detection of stress wave method, could be used in composite damage identification [66][67][68][69][70][71][72][73][74]. However, stress wave only generates under load. ...
The damage acoustic emission (AE) signal of refractories can be used to evaluate the classification and degree of the phase damage, which is different from the traditional damage location study in AE. In the analysis of AE signal, considering the nonuniformity of refractory inner structure, results of damage source inspection are usually affected by the signal transmission path. Therefore, the influence of transmitting path should be studied first. Piezoceramic material, capable of actuation and sensing, is commonly used to build AE probes. In this paper, two kinds of industrial refractories were selected as research objects and AE probes were chosen to generate emission acoustic signals. The relative amplitude attenuation coefficient and the relative energy attenuation coefficient were consequently designed as discriminant indicators to study the propagation characteristics of stress wave under different transmission paths and thicknesses. Results indicate that energy attenuation is higher than amplitude attenuation. Meanwhile, signal attenuation is principally proportional to transmission distance without directional selectivity. The above conclusions provide strong support to arrange AE sensors for nondestructive testing in refractory industry.
... [59][60][61]. In the literature, SVM has been used for fault diagnosis of rotating machine [62], gear fault diagnosis under variable conditions [63], source localization of acoustic emission [64], and defect diagnostics of Table 1 Training results for seven-class classification [57] Class No. of signals present ...
Composite materials are heterogeneous in nature and suffer from complex non-linear modes of failure, such as delamination, matrix crack, fiber-breakage, and voids, among others. The early detection of damage in composite structures, such as airplanes, is imperative to avoid catastrophic failure and tragic consequences. This paper reports on the use of machine learning techniques for the damage assessment (i.e., detection, quantification, and localization) of smart composite structures. The success of the machine learning paradigm for damage assessment depends on the representational capability of the discriminative features for the problems of interest. However, from a practical standpoint, it is not possible to define a global or superset of discriminative features that could discriminate between damaged and undamaged states of the structures, and simultaneously make a distinction between various modes of failures. In addition, one machine learning algorithm may show optimum performance for the discriminative features of a particular problem but fails for others. This article focuses on a review of discriminative features and the corresponding machine learning algorithms (both supervised and unsupervised), for various types of damage in smart composite structures.
... Chandrashekhar et al. [31] studied AE signals emitted in fatigue spectrum load tests. Noise elimination was conducted using un- supervised and supervised techniques. ...
The use of acoustic emission (AE) technique for damage diagnostic is typically challenging due to difficulties associated with discrimination of events that occur during different stages of damage that take place in a material or a structure. In this study, an unsupervised kernel fuzzy c-means pattern recognition analysis and the principal component method were utilized to categorize various damage stages in plain and steel fiber reinforced concrete specimens monitored by AE technique. Enhancement of the discrimination and characterization of damage mechanisms were achieved by processing time and frequency domain data. Both domains (time and frequency) were taken into account to propose new descriptors for crack classification purposes. A cluster of AE data in three classes of Kernel Fuzzy c-means (KFCM) was obtained. The clustered data was subsequently correlated with each particular damage stage for identifying the peak frequency range corresponding to the respective damage stages. Moreover, a novel quantitative technique called Spatial Intelligent b-value (SIb) Analysis was proposed to quantify damage for each stage.
... As a technique for structural health monitoring, AE can be used to detect micro-crack in structures under normal operation condition, unlike other nondestructive testing (NDT) techniques (for example ultrasonic), which require external input sources. This advantage makes it an attractive technique to monitor damage process on inservice reinforced structures with composite materials which has a complex failure mode due to the interaction of different materials [1][2][3][4][5]. Carbon fiber reinforced composite material, as a common used composite material, is increasingly applied in steel structure's rehabilitation due to its properties in terms of high strength to weight ratio and preferable corrosion resistance in last decade [6][7][8][9][10][11][12]. ...
Acoustic emission (AE) technique is a widely used CM technique. In this paper, AE technique was used to characterize the fatigue failure process for carbon fiber sheet (CFS) reinforced steel rod. The AE signals at the frequency band of 50–400 kHz are detected by using AE sensors mounted on the steel rod and analyzed by both parameter analysis and spectrum analysis in order to investigate the feasibility of using the AE technique to identify various failure modes during fatigue failure process for CFS reinforced steel rods. Tension-tension fatigue experiments were carried out on both CFS reinforced and unreinforced steel rods. Based on AE energy parameter analysis using wavelet decomposition method, failure process was initially divided into three stages for unreinforced specimen and seven stages for CFS reinforced specimen. The frequency contents within the frequency band of 50–400 kHz for various failure modes in each stage including crack initiation, developing, final rupture in steel rod and matrix crack, debonding in CFS were revealed by fast Fourier transform (FFT) method. Further wavelet transform (WT) analysis was performed to illustrate the sequences of the failure modes and main failure mode in each stage by the occurrence time and longest duration time, respectively. This work indicates that the proposed method is promising for distinguishing failure stages qualitatively and identifying failure modes quantitatively in CFS reinforced steel rods.
... A number of researchers have used UPR techniques such as FCM clustering with PCA (Principal Component Analysis), KSOM (Kohonen s Self-Organizing Map), K-Mean algorithm, Max-Min algorithm etc. Godin et al. used the pattern recognition technique to characterize the failure modes of the glass/polyester unidirectional specimens subjected to tensile loading [10] . Marec et al. used the Fuzzy C means with PCA to correlate each clusters to the failure mechanisms [14] , while Bhat et al. used the KSOM for the discrimination of the failure mechanisms and the noises present in it [15] . ...
Acoustic emission (AE) can be used for in situ structural health monitoring of the composite laminates. One of the main issues of AE is to characterize different dam-age mechanisms from the detected AE signals. In the present work, pure resin and GFRP composites laminates with different stacking sequences such as 0°, 90°, angle ply[±45°], cross-ply [0°/90°] are used to trigger different failure mechanisms when subjected to tensile test with AE monitoring. The study of failure mechanisms is facilitated by the choice of different oriented specimens in which one or two such mechanisms predominate. Range of peak frequencies in each orientation is investigated using FFT analysis. Fast Fourier Transform (FFT) enabled calculating the frequency content of each damage mechanism. Randomly selected hits from each range of peak frequencies for the specimens with different orientations subjected to tensile test with AE monitoring are analyzed using short time FFT (STFFT) analysis. STFFT analysis is used to highlight the possible failure mechanism associated with each signal. The predominance of failure modes in each orientation is useful in the study of discrimination of failure modes in composite laminates from AE data.
... AE is a real-time and in situ non-destructive testing method for health monitoring of the composite structures. Each AE signal originated from the active damage mechanisms has valuable information about the damages and can be considered as the acoustic signature of them [17][18][19][20]. Many researchers have already used AE to investigate damages occurred during propagation of delamination [21][22][23]. ...
... Signals collected from the pure resin, single fiber composites and unidirectional and cross-ply composites were classified with the Kohonen Self Organizing Map (KSOM) methods. It is widely recognized that pattern recognition technique can lead to good identification of AE data and a better understanding of the damage modes [13][14][15][16]. In particular, Gang Qi studied the fracture behavior of composite materials using wavelet based signal processing [17]. ...
The various failure mechanisms in bidirectional glass/epoxy laminates loaded in tension are identified using acoustic emission (AE) analysis. AE data recorded during the tensile testing of a single layer specimen are used to identify matrix cracking and fiber failure, while delami-nation signals are characterized using a two-layer specimen with a pre-induced defect. Parametric studies using AE count rate and cumulative counts allowed damage discrimination at different levels of loading and Fuzzy C-means clustering associated with principal component analysis were used to discriminate between failure mechanisms. The two above methods led to AE waveform selection: On selected waveforms, Fast Fourier Transform (FFT) enabled calculating the frequency content of each damage mechanism. Continuous wavelet transform allowed identifying frequency range and time history for failure modes, whilst noise content associated with the different failure modes was calculated and removed by discrete wavelet transform. Short Time FFT finally highlighted the possible failure mechanism associated with each signal.
... They obtain interesting results on unidirectional fiber-matrix and cross-ply composite AE data during tensile tests. Chandrashekhar et al. [22] have investigated AE signals generated by fatigue spectrum load tests on CFRP specimens. The combination of supervised and unsupervised methods eliminated noise, and AE signals were classified using a neural network technique. ...
In using acoustic emissions
(AE) for mechanical diagnostics, one major problem is the discrimination of events due to different types of damage occurring during loading of composite materials. In the present work, a procedure for the investigation of local damage in composite materials based on the analysis of the signals of Acoustic Emission
(AE) is presented. One of the remaining problems is the analysis of the AE signals in order to identify the most critical damage mechanisms. In this work, unsupervised pattern recognition
analyses (fuzzyc-means clustering) associated with a principal component analysis are the tools that are used for the classification of the monitored AE events. A cluster analysis of AE data is achieved and the resulting clusters are correlated to the damage mechanisms of the material under investigation. Time domain methods are used to determine new relevant descriptors to be introduced in the classification process in order to improve the characterization and the discrimination of the damage mechanisms. The results show that there is a good fitness between clustering groups and damage mechanisms. Also, AE with clustering procedure are as effective tools that provide a better discrimination of damage mechanisms in glass/polyester composite materials.
... Signals collected from the pure resin, single fiber composites, and unidirectional and cross-ply composites were classified with the Kohonen Self-Organising Map (KSOM) methods. It is widely recognized that pattern recognition technique can lead to good identification of AE data and a better understanding of the damage modes [13][14][15]. In particular, Qi [16] studied the fracture behavior of composite materials using wavelet based signal processing. ...
Acoustic emission (AE) is widely used to characterize damage occurring in composite materials: however, the discrimination between AE signatures due to different damage mechanisms is still an open issue. In this study, the various failure mechanisms in bidirectional glass/epoxy laminates subjected to uni-axial tension are identified using AE monitoring. AE data recorded during the tensile testing of a single-layer specimen are used to identify matrix cracking and fiber failure. In contrast, delamination signals are characterized using a two-layer specimen with a pre-induced defect, produced by artificially inserting a 10 mm wide Teflon tape in the middle portion of the two layers. Twelve-layer Glass fiber reinforced plastics laminates were also tested as a reference for the comparison of results. The procedure leading to signal discrimination involves a number of steps. First, Fuzzy C-means clustering associated with principal component analysis are used to discriminate between failure mechanisms, while parametric studies using AE count rate and cumulative counts allowed damage discrimination at various stages of loading. The two above methods led to AE waveform selection: on the selected waveforms, fast Fourier transform (FFT) enabled calculating the frequency content of each damage mechanism. Continuous wavelet transform (WT) allowed identifying frequency range and time history for failure modes in each signal, while noise content associated with the different failure modes is calculated and removed by discrete WT. Short time FFT (STFFT) finally highlighted the possible failure mechanism associated with each signal.
... They obtain interesting results on unidirectional fiber-matrix and cross-ply composite AE data during tensile tests. Chandrashekhar et al. [22] have investigated AE signals generated by fatigue spectrum load tests on CFRP specimens. The combination of supervised and unsupervised methods eliminated noise, and AE signals were classified using a neural network technique. ...
In using acoustic emissions (AEs) for mechanical diagnostics, one major problem is the discrimination of events due to different types of damage occurring during loading of composite materials. Unsupervised pattern recognition analyses (fuzzy c-means clustering) associated with a principal component analysis (PCA) are the tools that are used for the classification of the monitored AE events. Composites at different layups are used with the acoustic emission technique. A cluster analysis of AE data is achieved and the resulting clusters are correlated to the damage mechanisms of the material under investigation. Time domain methods are used to determine new relevant descriptors to be introduced in the classification process to improve the characterization and the discrimination of the damage mechanisms. The results show that there is a good fit between clustering groups and damage mechanisms. Additionally, AE with a clustering procedure are effective tools that provide a better discrimination of damage mechanisms in glass/polyester composite materials.
... The neuron corresponds to the classification of the input vector. Classification of AE data with SOM was applied by [20] with success in identifying different user-generated signals knowing a-priori the source mechanism. Usually, the number of neurons in SOM is much higher than the expected signal clusters. ...
Acoustic Emission (AE) is a promising technique for the damage detection and the real-time structural monitoring of composite lightweight structures; however data interpretation and discrimination among failure modes from AE data is difficult to be carried out without proper data processing techniques. In this paper, a neural-network based classification of AE signals from tensile tests of pultruded glass-fiber specimens is proposed. A self-organizing map is trained with AE data from one specimen; then the map is clustered with the k-means algorithm. The optimal number of clusters is chosen by a voting procedure that takes into account a number of quality indexes; then the clustered neural network is used to classify AE data from other specimen. Results have shown that the classifier built from a smooth specimen was able to correctly classify other specimens with the same and with a different material layup, and is capable of recognizing signals from notched specimens, thus providing interesting and encouraging indications in view of the application on real structures.
... Even if the damage mechanisms are quite well known (matrix cracking, aggregates fracture, splitting and debonding of the interface) their combinations and interactions are more complex to understand. For differentiating various damage mechanisms, some authors tried to relate each of the damage mechanisms with one range of frequencies [21,22] or amplitudes of AE signals [23][24][25]. ...
Delamination is the most common failure mode in composite materials, since it will result in the reduction of stiffness and can grow throughout other layers. Delamination is consisted of two main stages including initiation and propagation. Understanding the behavior of the material in these zones is very important, hence it has been thoroughly studied by different methods such as numerical methods, Acoustic Emission (AE), and modeling. Between these two regions initiation is a more vital stage in the delamination of the material. Once initiation occurs, which normally requires greater amount of force, cracks can easily propagate through the structure with little force and cause the failure of the structure. A better knowledge of initiation can lead to better design and production of stronger materials. Additionally, more knowledge about crack initiation and its internal microevents would help improve other parameters and result in higher strength against crack initiation. AE is a suitable method for in situ monitoring of damage in composite materials. In this study, AE was applied to test different glass/epoxy specimens which were loaded under mode I delamination. A function that combines AE and mechanical information is employed to investigate the initiation of delamination. Scanning electron microscope (SEM) was used to verify the results of this function. It is shown that this method is an appropriate technique to monitor the behavior of the initiation of delamination.
... Huguet and Godin et al. tracked the critical waveforms of different failure modes based on parametric based approach i.e. by using the amplitude parameter [13]. A. Marec used the Fuzzy C means clustering algorithm with PCA to correlate each clusters to the failure mechanisms [14]. C.R.L. Murthy in his paper have used the KSOM for the discrimination of the failure mechanisms and the noises present in it [15]. Chun-Gon Kim in his paper has studied the different failure modes using wavelet transform [16]. ...
In order to design structural components using composite materials a deep understanding of the material behaviour and its
failure mechanisms is necessary. To create a better understanding of the initiation, growth and interaction of the different
types of damage, damage monitoring during mechanical loading is very important. To this direction, AE is a powerful non destructive
technique for real time monitoring of damage development in materials and structures which has been used successfully for
the identification of damage mechanisms in composite structures under quasi static and dynamic-cycle loading. In this present
work, pure resin plate and GFRP composite laminates with stacking sequence of[00]6, are fabricated using Hand lay-up method. During the layup a Teflon tape of width 45mm is kept in the mid plane of the laminate
which serves as an initiator for delamination during loading. As per ASTM STD D552801DCB (Double Cantilever Beam) specimens
are cut out from the laminates and are subjected to tensile test in the transverse direction along with acoustic emission
monitoring. While loading, Markings are made on the sides of the specimen to track the crack front using a magnifying lens.
Parametric analysis is performed on the AE data obtained during crack propagation to discriminate the failure modes. Fast
Fourier Transform (FFT) enabled the calculation of frequency content of each damage mechanism. Further STFFT analysis is performed
on a portion of the waveforms representing the dominant frequency content pertaining to each damage mechanism.
KeywordsAcoustic emission–GFRP laminates–Mode I delamination–FFT and STFFT
Damage mechanisms in composite laminates are quite complex, and it is necessary to perceive their effects on the degradation of laminate mechanical properties. This work employs acoustic emission (AE) and digital image correlation (DIC) techniques to describe the evolution of intra/inter-laminar damage modes in the CFRP laminates under in-plane/out-of-plane loading conditions. In this study, laminates of stacking sequences [900]8, [450]8, [450/−450]2s, and [00]8 under tensile load are investigated to distinguish the intra-laminar damages like matrix cracking, fiber–matrix debond, and fiber breakage. Double cantilever beam, end notch flexure, and mixed-mode bending specimens are used to characterize delamination failure in the laminate. An unsupervised k-means clustering technique is used to classify the AE data based on peak frequency and amplitude. The surface displacement and strain data are evaluated using the DIC technique to understand the damage evolution in the laminates. Post failure analysis is carried out using a digital microscope, and fractography studies are used to identify and assign the damages to different AE clusters. This investigation yields a taxonomy of damage modes, their sequence of occurrence, and failure strains that can be used for structural health monitoring and progressive damage modeling of composite laminates.
Bridges are significant hubs in the U.S. national economy, facilitating the movement of goods and vehicles. The condition of bridges in the state of South Carolina is currently under scrutiny, especially in rural areas where most of the bridges were designed using outdated standards from the 1950 s. The weight of vehicles in recent years has increased significantly compared to the past. This has created an overloading problem. In addition, bridge performance decreases during their service life due to vehicle loads, material deterioration, and environmental erosion. Therefore, it is necessary to inspect and conduct load ratings on bridges to determine whether the bridges need to be posted. Due to recent advances in sensing technology and data analysis methods, nondestructive methods such as acoustic emission (AE) have been widely utilized in monitoring damage to the bridges. The objective of this paper is to explore the possibility of using AE sensors concurrently to determine vehicle loads on the bridges while monitoring bridge damage. A load determination method leveraging an improved ensemble artificial neural network (ANN) is proposed to analyze the AE data and estimate the load of the vehicle. The significance of this vehicle load determination method is that it has the potential to be paired with an AE damage monitoring system rather than using other instrumentation such as a weigh-in-motion (WIM) system. The proposed method has been tested on an experimental bridge component. The results suggest that the proposed model has an accuracy above 70 % in estimating the vehicle loads on the precast reinforced concrete (RC) flat slabs.
This thesis investigates the behaviour and failure of simple and complex
structures using the structural health monitoring system (SHM). The work focuses
on Acoustic Emission (AE) to detect, characterise and locate damage within
metallic and composites structures under a fatigue loading regime. The work was
divided into two main areas of research:
1. Damage Characterisation
Damage detection utilising AE was conducted through an extensive experimental
programme in large-scale carbon fibre composite structures. Different
assessment techniques were used to assess different damage mechanisms
within the structure under fatigue failure. The source mechanisms
characterisation in a large scale fatigue specimen was performed using a novel
parameter correction technique (PCT). This is a significant advance, offering (in
large scale structures) more reliable source characterisation.
2. Damage Localisation
Experimental investigations were undertaken to assess the novel AE location
technique proposed in this work in a variety of structures. The new technique,
known as Automatic Delta T mapping technique (Automatic DTM), provides an
accurate, easy to use, fast and reliable damage localisation technique.
Corrosion of steel reinforcement is a major deterioration mechanism in reinforced concrete structures that has led to highway bridge failures in the recent past. While prestressed and posttensioned concrete structures offer natural protection to the reinforcement by limiting crack development, the high stresses in prestressing strands and their geometry facilitate crevice corrosion that accelerates corrosion deterioration and may increase the risk of sudden failure. Traditional methods for corrosion detection, such as visual inspection and electrochemical measurements, are local, time-consuming, and may not be feasible in posttensioned concrete structures. This chapter summarizes recent efforts that utilize the acoustic emission technique for corrosion detection in passively reinforced, prestressed, and posttensioned concrete structures. Acoustic emission is sensitive to ongoing damage and can detect at both the micro and the macro level. The method is nonintrusive and can enable global assessment of the structural condition. The methodology is reviewed, and the results of accelerated corrosion experiments on specimens with different sizes and configurations are reported. In these studies, acoustic emission results are validated using electrochemical results and micrographs from scanning electron microscopy. A newly developed acoustic emission–based corrosion classification chart is also presented.
A non-destructive testing (NDT) system based on a superconducting quantum interference device (SQUID) is used for detecting the electromagnetic properties of three-dimensional (3D) braided composites. Samples with artificial defects are detected via artificial neural network and the flux imaging methods; the results compared well. Experimental results indicate that the SQUID-based NDT technique is far more advanced and practical for three-dimensional (3D) braided composites.
In the present work damage was monitored in flax/epoxy quasi unidirectional woven laminates. Several plates with different lay-up configurations: [0°]8,[0°,90°]2s,[0°,90°,+45°,-45°]s and [+45,-45]2s were prepared in autoclave, then the damage was monitored during the tensile test using acoustic emission technique. The tensile tests show that these composites offer good mechanical properties. The acoustic emission diagrams allowed us to follow the evolution of damage and to identify several parameters: Energy, damage threshold and the number of events, however the correlation between the stress-strain curves and AE results don't show a direct relationship between the two damage indicators (E&AE): this suggests that the shape of the strain-strain curves is due predominately to another factor than AE events.
Acoustic Emission (AE) can be used to discriminate the different types of damage occurring in a constrained material. However, in industrial surrounding the main problem associated with data analysis is the discrimination between the noises and the acoustic emission signals. The goal of our paper was to differentiate between air flow and AE signals produced by the material during thermal shock of a refractory material in a pseudo-industrial surrounding with multiparameter analysis. The clustering of AE data corrupted by noise was successfully achieved by the k-means method.
The increasing popularity of structural health monitoring has brought with it a growing need for automated data management and data analysis tools. Of great importance are filters that can systematically detect unwanted signals in acoustic emission datasets. This study presents a semi-supervised data mining scheme that detects data belonging to unfamiliar distributions. This type of outlier detection scheme is useful detecting the presence of new acoustic emission sources, given a training dataset of unwanted signals. In addition to classifying new observations (herein referred to as “outliers”) within a dataset, the scheme generates a decision tree that classifies sub-clusters within the outlier context set. The obtained tree can be interpreted as a series of characterization rules for newly-observed data, and they can potentially describe the basic structure of different modes within the outlier distribution. The data mining scheme is first validated on a synthetic dataset, and an attempt is made to confirm the algorithms’ ability to discriminate outlier acoustic emission sources from a controlled pencil-lead-break experiment. Finally, the scheme is applied to data from two fatigue crack-growth steel specimens, where it is shown that extracted rules can adequately describe crack-growth related acoustic emission sources while filtering out background “noise.” Results show promising performance in filter generation, thereby allowing analysts to extract, characterize, and focus only on meaningful signals.
The use of Acoustic Emission (AE) as a Structural Health Monitoring (SHM) technique is very attractive thanks to its ability to detect not only damage sources in real-time but also to locate them.
To demonstrate the AE capabilities on known damage modes, a carbon fibre panel was manufactured with cut fibres in a central location and subjected to fatigue loading to promote matrix cracking. Subsequently, a delamination was created within the panel using an impact load, and the test was continued.
AE signals were located within the crack area in the first part of the test. After impact, AE signals were detected from both areas under fatigue loading; signals from this area were located and used for further analysis with the neural network technique.
The application of an unsupervised neural network based classification technique successfully separated two damage mechanisms, related to matrix cracking and delamination. The results obtained allowed a more detailed understanding of such sources of AE in carbon fibre laminates.
The purpose of this study is to investigate the damage mechanisms in UHMWPE/LDPE laminated by Acoustic Emission (AE) technique. Model specimens are fabricated to obtain expected damage mechanisms during tensile testing. Then, relationship among AE descriptors is studied by hierarchical cluster analysis, and AE signals are classified by k-means algorithm. Finally, an Artificial Neural Network (ANN) is created and trained by various optimal algorithms to identify damage mechanisms. The results reveal that typical damage mechanisms in PE self-reinforced composite can be classified in terms of the similarity between AE signals and identified by a well trained ANN. © 2012 Binary Information Press & Textile Bioengineering and Informatics Society.
Thermoplastic self-reinforced polyethylene (PE/PE) composites were tested under quasi-static tensile load and the failure processes weremonitored by Acoustic Emission (AE) technique. The AE signals were collected and clustered by Unsupervised Pattern Recognition (UPR) scheme. The initiation and progression of the damage mechanisms in the composites can then be reviewed by the cumulative AE hits of each cluster versus strain curves. But the labeling of each cluster is crucial to the failure analysis. The paper focuses on this correlating between the obtained clusters and their specific damage modes. This was carried out by waveform visualization and Fast Fourier Transform analysis. Pure resin and fiber bundles were tested to assist in the labeling of signal classes in the composites (90°, 0° and [±45°] specimens). Typical waveforms of matrix cracking, fiber-matrix debonding, fiber fracture and fiber pullout were indentified respectively. The evolution process of various damage mechanisms in the composites revealed that the correlating method was effective. An objective and repeatable analytical procedure is established for the investigation of progressive failure mechanisms in the thermoplastic composites.
This paper provides an overview of research in Aerospace structures in India by premier academic institutions and research laboratories. The research programs have been closely influenced by the aerospace flight vehicle programs in India. Research funding was available based on the design requirements and the Indian researchers covered a wade spectrum of research areas. The major sponsor of aerospace structures research has been Aeronautics R&D Board and in space related research the major sponsor has been Indian Space Research organization. This paper highlights some of the major contributions. Copyright © 2009 by the American Institute of Aeronautics and Astronautics, Inc.
Fiber reinforced composite materials are used as structural material in airplanes because of their high specific stiffness and strength. When composite materials are subjected to mechanical loading, it leads to many types of failures such as matrix cracking, debonding, delamination, and fiber breakage. To create a better understanding of the initiation, growth and interaction of the different types of damage, damage monitoring during mechanical loading is very important. Acoustic emission is a suitable technique for the detection of a wide range of micro-structural failures in composite materials. In this present paper, GFRP composite laminates with different stacking sequences such as [0°]4, [90°]4, angle ply [± 45°]4 and cross ply [0°/90°]4 are used to trigger different failure mechanisms when subjected to flexural test (three-point bending) with AE monitoring. Discrimination of the failure modes are done based on the predominance of the failure modes in each orientation. Parametric plots are used to discriminate the modes of fracture within the laminates. Range of frequency content in each orientation is investigated using fast Fourier transform (FFT) analysis. FFT enables calculating the frequency content of each damage mechanism. Short Time FFT highlights the possible failure mechanisms associated with each signal. Continuous wavelet transform allowed identifying frequency range and time history for failure modes in each signal. The predominance of failure modes in each orientation is used as a key in the study of discrimination of failure modes in composite laminates.
The study concerns with classification of acoustic emission signals in composite laminates using support vector machine (SVM). Wavelet packet analysis is performed initially to extract the features and to reduce the dimensionality of original data features. The SVM classifiers are trained with a subset of the experimental data for known fault conditions and are tested using the remaining set of data. The result shows that muti-class SVM produces promising results and has potential for use in AE signal classification.
Acoustic emission (AE) signals collected from thermoplastic self-reinforced polyethylene composites ultra-high molecular weight polyethylene fibre reinforced low-density polyethylene (UHMWPE/LDPE) under quasi-static tensile load were clustered and identified by unsupervised pattern recognition (UPR) and supervised pattern recognition (SPR) techniques in order to clarify various damage modes in the composites. The purpose was to find an easy way to separate a set of data with a large number of unknown AE signals into several classes attributed to a specific damage mode each. UPR techniques were utilised first to classify the AE signals from simple lay-up laminate specimens automatically and mathematically. Different damage modes were identified and a physical validation was carried out by the scanning electron microscope (SEM) technique. Damage investigation of the specimen according to the clustering results showed reasonable results. Therefore, the labelling data set consisting of signals from different damage modes was used as the reference for a SPR system. A large number of AE signals from quasi-isotropic laminates were then identified by the supervised method. It showed good results, which were also supported by the SEM examination. A reliable and convenient procedure was established for a good identification of large number of unknown AE data for the UHMWPE/LDPE composites. This methodology is promising for any other fibre reinforced composites in the field of damage mechanisms analysis.
The objective of present study is to classify and identify damage mechanisms in polyethylene(PE) self-reinforced composites by acoustic emission (AE) technique. Model specimens including LDPE resin, [90°]laminate, single fiber composite, fiber bundle composite, and [±45°] laminates are fabricated to obtain expected damage mechanisms during tensile testing. First, mechanical behaviors and corresponding AE response of model specimens are studied to validate damage mechanisms in UHMWPE/LDPE laminates. Second, relationship among AE descriptors is investigated by hierarchical cluster analysis, and AE signals are classified by k-means cluster analysis. Correlations between damage mechanisms and AE are established in terms of amplitude, duration, and peak frequency of AE signals. Finally, an artificial neural network is created and trained by various optimal algorithms to identify damage mechanisms. The results reveal that typical damage mechanisms in PE self-reinforced composite can be classified in terms of the similarity between AE signals and identified by trained artificial neural network. POLYM. COMPOS., 2011. © 2011 Society of Plastics Engineers
An objective analytical procedure for the investigation of damage mechanisms in the thermoplastic self-reinforced polyethylene
(UHMWPE/PE) composites under quasi-static tensile load has been established, using Unsupervised Pattern Recognition (UPR)
technique for the clustering task of Acoustic Emission (AE) signals. Focus is on the correlating between the obtained classes
and their specific damage mechanisms. This was carried out by waveform visualization and Fast Fourier Transform analysis.
Pure resin and fiber bundles were tested to collect typical waveforms of matrix cracking and fiber fracture respectively,
in order to label the signal classes in the composites. The evolution process of various damage mechanisms in the composites
revealed that the correlating method was effective. The AE characteristics of different damage modes found out in this study
can be used as the reference for identifying unknown AE signals in the UHMWPE/PE composites. The established procedure is
also potential in the investigation of failure mechanisms for composite materials with UPR technique.
Neural networks have been widely used for many applications. One of the applications is forecasting. Many studies have proven
that neural networks can provide good accuracy on forecasting future data with over than 80% accuracy. In this study, neural
network is used to predict bearing defects. Two learning tasks, function approximation and pattern recognition, were used
for detection and monitoring of defects in ball bearing. Given five categories of bearing defect, the neural networks have
successfully proven the ability to distinguish one defect over the other with high accuracy. Acoustic emission (AE) was used
as a measurement in this study. AE is defined as transient waves generated from a rapid release of strain energy by deformation
or damage or on the surface of a material (1–3). The AE waves can provide information about bearing condition. Maximum amplitude and AE counts were used as the basis for
detection.
KeywordsPredict-Ball bearing defects-Neural network
The effects of hydrolytic ageing on the acoustic emission signature of damage mechanisms occurring during tensile tests on a glass fibre reinforced polyester composite are investigated. These results are used to test the validity of a Kohonen’s map established with AE data collected from an unaged specimen. This method enabled identification and classification of AE signals belonging to two different failure modes. The Kohonen self-organising map, trained on unaged specimens, is a valuable tool in assessment of damage type on the aged specimens.
The first production use of practical acoustic emission (AE) on aircraft was on the F-111, where more than 240 U.S. and Australian aircraft have been successfully monitored during cold proof testing since 1987. A key factor was the design of an instrument which automatically configured itself at power-on, so that aircraft testing could take place in a highly efficient and reliable manner. The second production use of AE on aircraft was on the VC-10, where an entire fleet of 22 aircraft was monitored 40 times during pneumatic proof pressurization. The monitoring of this large transport required anywhere from 282 to 313 narrowband AE sensors per aircraft. In both aircraft types, numerous significant defects were discovered through AE-based nondestructive testing. Finally, we present on-going in-flight AE research. This new research uses digital waveform processing of wideband AE signals and offers the potential adding new and complementary capabilities to classical, narrowband AE.
A structural article representative of a large full-production military aircraft wing was subjected to flight-by-flight load spectra to initiate and propagate fatigue cracks in 17 selected fastener holes in eight test areas. One-hundred unit flights were applied to the wing to initiate crack growth at sharp notch sawcuts and existing cracks, followed by 2,000 unit flights to propagate the cracks. Seven test areas were monitored for acoustic emission (AE) using a 32-channel AE flaw locator system and triangulated sensor arrays. After the tests, the test holes were removed from the structure for fractographic analysis of crack growth. Correlations between AE and crack growth for several of the test holes are discussed.
Acoustic emission source location techniques were successfully used on a production-size aircraft wing fatigue test article to monitor crack growths in the range from 0.25 mm to 1.6 mm per load cycle. The AE data showed good correlation with the crack length data. Analyses of these correlations show that AE monitoring has the potential for being used to determine crack length over this range of crack growth rates. The fatigue test article was constructed of 7075-T6511 aluminum alloy. The test period lasted 14 days and the results demonstrated that AE has application to aircraft structures.
Significant portions of the F-111 aircraft were fabricated of D6AC steel, which is now known to have a fairly small critical crack size. To prove structural flight-worthiness the Air Force built a chamber at McClellan AFB, where all F-111 aircraft are periodically chilled to 40°C and stressed to 7.3g and3.0g. Recently the chamber was modernized, and Physical Acoustics Corporation was selected to supply an acoustic emission system to locate any sources of structural failure. The new F-111 Cold Proof Test Station Acoustic Emission Monitoring System has several innovative features, including a colour CRT which displays the exact location of AE events in real time on simultaneous overhead and side views of the F-111. The events are coloured green, yellow and red according to their severity, as calculated from their amplitude and energy. A monochromatic CRT is used concurrently with the colour CRT to display severity information on AE events which only arrive at one sensor. Alarms also audibly alert the operator to crucial events, using two tones to distinguish the degree of severity.
Safety and reliability are prime concerns in aircraft performance due to the involved costs and risk to lives. Despite the best efforts in design methodology, quality evaluation in production and structural integrity assessment in-service, attainment of one hundred percent safety through development and use of a suitable in-flight health monitoring system is still a farfetched goal. And, evolution of such a system requires, first, identification of an appropriate Technique and next its adoption to meet the challenges posed by newer materials (advanced composites), complex structures and the flight environment. In fact, a quick survey of the available Non-Destructive Evaluation (NDE) techniques suggests Acoustic Emission (AE) as the only available method. High merit in itself could be a weakness - Noise is the worst enemy of AE. So, while difficulties are posed due to the insufficient understanding of the basic behavior of composites, growth and interaction of defects and damage under a specified load condition, high in-flight noise further complicates the issue making the developmental task apparently formidable and challenging. Development of an in-flight monitoring system based on AE to function as an early warning system needs addressing three aspects, viz., the first, discrimination of AE signals from noise data, the second, extraction of required information from AE signals for identification of sources (source characterization) and quantification of its growth, and the third, automation of the entire process. And, a quick assessment of the aspects involved suggests that Artificial Neural Networks (ANN) are ideally suited for solving such a complex problem. A review of the available open literature while indicates a number of investigations carried out using noise elimination and source characterization methods such as frequency filtering and statistical pattern recognition but shows only sporadic attempts using ANN. This may probably be due to the complex nature of the problem involving investigation of a large number of influencing parameters, amount of effort and time to be invested, and facilities required and multi-disciplinary nature of the problem. Hence as stated in the foregoing, the need for such a study cannot be over emphasized. Thus, this thesis is an attempt addressing the issue of analysis and automation of complex sets of AE data such as AE signals mixed with in-flight noise thus forming the first step towards in-flight monitoring using AE. An ANN can in fact replace the traditional algorithmic approaches used in the past. ANN in general are model free estimators and derive their computational efficiency due to large connectivity, massive parallelism, non-linear analog response and learning capabilities. They are better suited than the conventional methods (statistical pattern recognition methods) due to their characteristics such as classification, pattern matching, learning, generalization, fault tolerance and distributed memory and their ability to process unstructured data sets which may be carrying incomplete information at times and hence chosen as the tool. Further, in the current context, the set of investigations undertaken were in the absence of sufficient a priori information and hence clustering of signals generated by AE sources through self-organizing maps is more appropriate. Thus, in the investigations carried out under the scope of this thesis, at first a hybrid network named "NAEDA" (Neural network for Acoustic Emission Data Analysis) using Kohonen self-organizing feature map (KSOM) and multi-layer perceptron (MLP) that learns on back propagation learning rule was specifically developed with innovative data processing techniques built into the network. However, for accurate pattern recognition, multi-layer back propagation NN needed to be trained with source and noise clusters as input data. Thus, in addition to optimizing the network architecture and training parameters, preprocessing of input data to the network and multi-class clustering and classification proved to be the corner stones in obtaining excellent identification accuracy. Next, in-flight noise environment of an aircraft was generated off line through carefully designed simulation experiments carried out in the laboratory (Ex: EMI, friction, fretting and other mechanical and hydraulic phenomena) based on the in-flight noise survey carried out by earlier investigators. From these experiments data was acquired and classified into their respective classes through MLP. Further, these noises were mixed together and clustered through KSOM and then classified into their respective clusters through MLP resulting in an accuracy of 95%- 100% Subsequently, to evaluate the utility of NAEDA for source classification and characterization, carbon fiber reinforced plastic (CFRP) specimens were subjected to spectrum loading simulating typical in-flight load and AE signals were acquired continuously up to a maximum of three designed lives and in some cases up to failure. Further, AE signals with similar characteristics were grouped into individual clusters through self-organizing map and labeled as belonging to appropriate failure modes, there by generating the class configuration. Then MLP was trained with this class information, which resulted in automatic identification and classification of failure modes with an accuracy of 95% - 100%. In addition, extraneous noise generated during the experiments was acquired and classified so as to evaluate the presence or absence of such data in the AE data acquired from the CFRP specimens. In the next stage, noise and signals were mixed together at random and were reclassified into their respective classes through supervised training of multi-layer back propagation NN. Initially only noise was discriminated from the AE signals from CFRP failure modes and subsequently both noise discrimination and failure mode identification and classification was carried out resulting in an accuracy of 95% - 100% in most of the cases. Further, extraneous signals mentioned above were classified which indicated the presence of such signals in the AE signals obtained from the CFRP specimen. Thus, having established the basis for noise identification and AE source classification and characterization, two specific examples were considered to evaluate the utility and efficiency of NAEDA. In the first, with the postulation that different basic failure modes in composites have unique AE signatures, the difference in damage generation and progression can be clearly characterized under different loading conditions. To examine this, static compression tests were conducted on a different set of CFRP specimens till failure with continuous AE monitoring and the resulting AE signals were classified through already trained NAEDA. The results obtained shows that the total number of signals obtained were very less when compared to fatigue tests and the specimens failed with hardly any damage growth. Further, NAEDA was able to discriminate the"noise and failure modes in CFRP specimen with the same degree of accuracy with which it has classified such signals obtained from fatigue tests. In the second example, with the same postulate of unique AE signatures for different failure modes, the differences in the complexion of the damage growth and progression should become clearly evident when one considers specimens with different lay up sequences. To examine this, the data was reclassified on the basis of differences in lay up sequences from specimens subjected to fatigue. The results obtained clearly confirmed the postulation. As can be seen from the summary of the work presented in the foregoing paragraphs, the investigations undertaken within the scope of this thesis involve elaborate experimentation, development of tools, acquisition of extensive data and analysis. Never the less, the results obtained were commensurate with the efforts and have been fruitful. Of the useful results that have been obtained, to state in specific, the first is, discrimination of simulated noise sources achieved with significant success but for some overlapping which is not of major concern as far as noises are concerned. Therefore they are grouped into required number of clusters so as to achieve better classification through supervised NN. This proved to be an innovative measure in supervised classification through back propagation NN. The second is the damage characterization in CFRP specimens, which involved imaginative data processing techniques that proved their worth in terms of optimization of various training parameters and resulted in accurate identification through clustering. Labeling of clusters is made possible by marking each signal starting from clustering to final classification through supervised neural network and is achieved through phenomenological correlation combined with ultrasonic imaging. Most rewarding of all is the identification of failure modes (AE signals) mixed in noise into their respective classes. This is a direct consequence of innovative data processing, multi-class clustering and flexibility of grouping various noise signals into suitable number of clusters. Thus, the results obtained and presented in this thesis on NN approach to AE signal analysis clearly establishes the fact that methods and procedures developed can automate detection and identification of failure modes in CFRP composites under hostile environment, which could lead to the development of an in-flight monitoring system.
Characterization of noise generated in HS 748 aircraft in-flight using acoustic emission
- John Sk
- R Sudha
- Murthy
- Crl
John SK, Sudha R, Murthy CRL. Characterization of noise generated in HS 748 aircraft in-flight using acoustic emission. J Aeronaut Soc India 1994.
Practical AE methodology for use on aircraft Acoustic Emission; Standards and Technol-ogy Update, ASTM STP 1353
- Carlyle Jm Bodiline Hl
- Henley
- Dawes Rl Ss
- Hill
- Evk
Carlyle JM, Bodiline HL, Henley SS, Dawes RL, Hill EVK. Practical AE methodology for use on aircraft. In: Vah-aviolos SJ, editor. Acoustic Emission; Standards and Technol-ogy Update, ASTM STP 1353. West Conshohocken, PA: American Society for Testing and Materials; 1999. p. 191– 205.
Acoustic emission monitoring of aircraft noises during a ground test
- R Gulshion
- Crl Murthy
Gulshion R, Murthy CRL. Acoustic emission monitoring of aircraft noises during a ground test. In: Hallai C, Kulcsar P, editors. Proc. of the 13th World conference on Non-Destructive, 1992.
In-flight crack detection system for the C-135 lower center wing skin
- Mizell
- Wilson Me
- Jr
Mizell ME, Wilson Jr TL. In-flight crack detection system for the C-135 lower center wing skin. ISA ASI 76240 1976:259–68.
Evaluation of crack initiation and growth from simulated flight environment by AE signal analysis. ME Project
- Muthu Krishan
Muthu Krishan. Evaluation of crack initiation and growth from simulated flight environment by AE signal analysis. ME Project, Department of Aerospace Engineering, IISc., Bangalore, India, 1995.
Inspecting aging aircrafts
- Carlyle Jm Vahaviolos
- Sj
Carlyle JM, Vahaviolos SJ. Inspecting aging aircrafts. Quality 1989; No. 11. Reprinted in Physical Acoustics InstitutionÕs, 1989, selected papers publication.
Some studies on acoustic emission for in-flight monitoring of critical aircraft components
- Ramakrishna R Hv Gulshion
- Bhat Mr
- Murthy
- Crl
Gulshion R, Ramakrishna HV, Bhat MR, Murthy CRL. Some studies on acoustic emission for in-flight monitoring of critical aircraft components. J Nondestruct Eval 1990;10(3).
Acoustic emission––monitoring fatigue
- Baily Cd Wm
Baily CD, Pless WM. Acoustic emission––monitoring fatigue
Noise suppression in acoustic emission testing: an approach
- Crl Murhty
- K K Brhama
- N K Badri
- M R Bhat
Murhty CRL, Brhama KK, Badri NK, Bhat MR. Noise suppression in acoustic emission testing: an approach. In: Proc. of the Conference on Metallography and NDT. Interactions in Materials Evaluation, BARC, 1986. p. IIIA-a.8.
Some studies on acoustic emission for in-flight monitoring of critical aircraft components
- Gulshion
Characterization of noise generated in HS 748 aircraft in-flight using acoustic emission
- John
Acoustic emission--monitoring fatigue cracks in aircraft structure
- C D Baily
- W M Pless
Inspecting aging aircrafts
- J M Carlyle
- S J Vahaviolos