Showing papers in "Journal of Nondestructive Evaluation in 2020"

Physics-Informed Neural Network for Ultrasound Nondestructive Quantification of Surface Breaking Cracks

Abstract: We introduce an optimized physics-informed neural network (PINN) trained to solve the problem of identifying and characterizing a surface breaking crack in a metal plate. PINNs are neural networks that can combine data and physics in the learning process by adding the residuals of a system of partial differential equations to the loss function. Our PINNs is supervised with realistic ultrasonic surface acoustic wave data acquired at a frequency of 5 MHz. The ultrasonic surface wave data is represented as a deformation on the top surface of a metal plate, measured by using the method of laser vibrometry. The PINN is physically informed by the acoustic wave equation and its convergence is sped up using adaptive activation functions. The adaptive activation function uses a trainable hyperparameter, which is optimized to achieve the best performance of the network. The adaptive activation function changes dynamically, involved in the optimization process. The usage of the adaptive activation function significantly improves the convergence, evidently observed in the current study. We use PINNs to estimate the speed of sound of the metal plate, which we do with an error of 1%, and then, by allowing the speed of sound to be space dependent, we identify and characterize the crack as the positions where the speed of sound has decreased. Our study also shows the effect of sub-sampling of the data on the sensitivity of sound speed estimates. More broadly, the resulting model shows a promising deep neural network model for ill-posed inverse problems.

Overview of Researches on the Nondestructive Testing Method of Metal Magnetic Memory: Status and Challenges

Abstract: More than 20 years of research progress regarding the nondestructive testing method of metal magnetic memory is reviewed and summarized in detail. Consequently, this overview is selective, covering what we feel are the most important trends of experimental phenomena, mechanism explanations, quantitative theories, simulations, testing, evaluation and application. From analyzing the current state of research on the method of metal magnetic memory, some key problems and future developmental trends are proposed. Although the research on magnetic memory method has made great progress, the practical application still faces problems such as complex influencing factors and less quantitative research. In the future, for magnetic memory method, it is necessary to strengthen the microscopic observations of magnetic domains, experiments of magnetomechanical constitutive, establishment of quantitative models, modeling of complex influencing factors, and the study of identification, inversion and criteria. In addition, the combination of other non-destructive testing methods can greatly improve the practical application of the magnetic memory method.

On the Use of X-ray Computed Tomography in Assessment of 3D-Printed Components

Abstract: Technical advantages of additive manufacturing (AM) have drawn great attention over the past few years. This cost-effective manufacturing process proved its potential applications in a wide range of fields. Although AM techniques (known as 3D printing) are able to fabricate geometrically complex components, it is necessary to evaluate internal and external dimensions of the printed parts. In this context, x-ray computed tomography (CT) as a nondestructive evaluation technique has been utilized. Indeed, CT can be used for geometric analysis, defects detection, quantitative comparison, structural quantification and porosity analysis. In the current study, we present a brief review of 3D printing processes and evolution of CT technology. Moreover, applications of CT in assessment of 3D-printed components are explained in detail. Although CT has been used in academic and industrial researches, abilities of this inspection method are not yet fully documented for precision engineering applications. In this work, usage of this technique in study of printed components are categorized in four subdomains and discussed. The documented data proved that CT is an appropriate non-contact technique for technical evaluation of various printed parts. As usage of CT in assessment of printed parts is still evolving, the limitations, challenges and future perspective are outlined.

Aluminum Casting Inspection Using Deep Learning: A Method Based on Convolutional Neural Networks

Abstract: Castings produced for the automotive industry are considered important components for overall roadworthiness. To ensure the safety of construction, it is necessary to check every part thoroughly using non-destructive testing. X-ray testing rapidly became the accepted way of controlling the quality of die-cast pieces. In this paper, we present a Convolutional Neural Network (CNN) for defect detection in castings. In order to train the CNN model, a large dataset is necessary. We build the dataset by using synthetic defects. They are simulated using 3D ellipsoidal models and Generative Adversarial Networks (GAN). We compare different portions of ellipsoidal/GAN defects in the training subset. In our experiments, the use of GAN defects does not play a relevant role in this solution. However, ellipsoidal defects helped to achieve better performance. Ellipsoidal defects from any size and orientation could be superimposed onto real X-ray images in any location. In addition, we tested several CNN configurations, the best one, that we call Xnet-II, has 30 layers and more than 1,350,000 parameters. It has been trained using a dataset with around 640,000 patches containing 50% of ellipsoidal defects and 50% of real background captured from different casting types. The model was tested using sliding-windows methodology on whole X-ray images achieving promising results (mPA = 0.7102): the model was able to detect real defects from different casting types. We believe that the methodology presented could be used in similar projects that have to deal with automated detection of defects.

A Review of Wire Rope Detection Methods, Sensors and Signal Processing Techniques

Abstract: Wire rope inspection by nondestructive testing methods, sensors and signal processing techniques are mainly reviewed in this paper. Owing to the difference of physical mechanism and testing principles, magnetic flux leakage, eddy current, acoustic emission and ultrasonic guide wave testing as well as other inspection methods for steel wire rope are summarized. Then, the commonly and frequently used testing sensors of inductive coil, hall element, magnetoresistive sensors and others are compared in the perspective of their corresponding operating principles, development situation, advantages and disadvantages. Furthermore, signal processing techniques including the signal filtering techniques such as the time and frequency analysis methods, quantitative data processing methods such as the machine learning and defect classification are studied. Finally, the challenges and future developing trends of wire rope inspection in practical applications are discussed.

A Review of the Metal Magnetic Memory Method

Abstract: Metal magnetic memory (MMM) method is a non-destructive testing (NDT) technology which has potentials to detect early damage. A review is presented in this paper about the development of this method, including the theoretical studies of the magnetic/stress coupling effect, factors influencing the detection signals, the criteria for judging the damage state and defect identification. Directions of future research are also discussed.

Using Deep Learning for Defect Classification on a Small Weld X-ray Image Dataset

Abstract: This document provides a comparative evaluation of the performance of a deep learning network for different combinations of parameters and hyper-parameters. Although there are numerous studies that report on performance in deep learning networks for ordinary data sets, their performance on small data sets is much less evaluated. The objective of this work is to demonstrate that such a challenging small data set, such as a welding X-ray image data set, can be trained and evaluated obtaining high precision and that it is possible thanks to data augmentation. In fact, this article shows that data augmentation, also a typical technique in any learning process on a large data set, plus that two image channels, such as channels B (blue) and G (green), both are replaced by the Canny edge map and a binary image provided by an adaptive Gaussian threshold, respectively, gives to the network a 3% increase in accuracy, approximately. In summary, the objective of this work is to present the methodology used and the results obtained to estimate the classification accuracy of three main classes of welding defects obtained on a small set of welding X-ray image data.

A GMDH Predictive Model to Predict Rock Material Strength Using Three Non-destructive Tests

Abstract: The uniaxial compressive strength (UCS) is considered as a significant parameter related to rock material in design of geotechnical structures connected to the rock mass. Determining UCS values in laboratory is costly and time consuming, hence, its indirect determination through use of rock index tests is of a great interest and advantage. This study presents a prediction process of the UCS values through the use of three non-destructive tests i.e., p-wave velocity, Schmidt hammer and density. This process was done by developing an intelligent predictive technique namely the group method of data handling (GMDH). Before constructing intelligence system, a series of experimental equations were proposed using three non-destructive tests. The results showed that there is a need to propose new model with taking advantages of all three non-destructive tests results. Then, several GMDH models were built through the use of various parametric studies on the most effective GMDH factors. For comparison purposes, an artificial neural network (ANN) was also modelled to predict rock strength. The obtained results of the ANN and GMDH were assessed based on system error and coefficient of determination values. The results confirmed that the proposed GMDH model is an applicable, powerful, and practical intelligence system that is able to provide an acceptable accuracy level for predicting rock strength.

A study of the automated eddy current detection of cracks in steel plates

Abstract: Applying life estimation approaches to determine in-service life of structures and plan the inspection schedules accordingly are becoming acceptable safety design procedures in aerospace. However, these design systems shall be fed with reliable parameters related to material properties, loading conditions and defect characteristics. In this context, the role of non-destructive (NDT) testing reliability is of high importance in detecting and sizing defects. Eddy current test (ECT) is an electromagnetic NDT method frequently used to inspect tiny surface fatigue cracks in sensitive industries. Owing to the new advances in robotic technologies, there is a trend to integrate the ECT into automated systems to perform NDT inspections more efficiently. In fact, ECT can be effectively automated as to increase the coverage, repeatability and scanning speed. The reliability of ECT scanning, however, should be thoroughly investigated and compared to conventional modes of applications to obtain a better understanding of the advantages and shortcomings related to this technique. In this contribution, a series of manual and automated ECT tests are carried out on a set of samples using a split-D reflection differential surface probe. The study investigates the level of noise recorded in each technique and discuss its dependency on different parameters, such as surface roughness and frequency. Afterwards, a description of the effect of crack orientation on ECT signal amplitude is provided through experimental tests and finite element simulations. Finally, the reliability of each ECT technique is investigated by means of probability of detection (POD) curves. POD parameters are then extracted and compared to examine the effect of scanning index, frequency and automation on detection reliability.

Stress-dependent magnetic flux leakage: finite element modelling simulations versus experiments

Abstract: Assessing the effect of defect induced stresses on magnetic flux leakage (MFL) signals is a complicated task due to nonlinear magnetomechanical coupling. To facilitate the analysis, a multi-physics finite elemental simulation model is proposed based on magnetomechanical theory. The model works by quasi-statically computing the stress distribution in the specimen, which is then inherited to solve the nonlinear magnetic problem dynamically. The converged solution allows identification and extraction of the MFL signal induced by the defect along the sensor scanning line. Experiments are conducted on an AISI 1045 steel specimen, i.e. a dog-bone shaped rod with a cylindrical square-notch defect. The experiments confirm the validity of the proposed model that predicted a linear dependency of the peak-to-peak amplitude of the normalized MFL signal on applied stress. Besides identifying the effect of stress on the induced MFL signal, the proposed model is also suitable for solving the inverse problem of sizing the defects when stress is involved.

A New Ultrasonic Amplitude Tomography Approach, with Validation on Masonry Tuff Blocks

Abstract: Despite sonic and ultrasonic tomography are major NDT techniques, the resolution and the capability of these tests could be substantially improved; this, especially for masonry structures Moreover, although waves amplitude attenuation is strictly related to the internal features and to the defects, some experimental and theoretical issues obstruct practical applications of amplitude tomography for civil structures Indeed, almost exclusively travel time tomography is employed Here, a new approach for ultrasonic amplitude tomography called Standardized Amplitude Tomography (SAT) is proposed This approach is based on a suitable formulation of the theoretical model to be implemented for the tomographic reconstruction of the internal features of materials and on the employ of a special experimental tool SAT approach is primarily meant for application to masonry structures, but its principle is applicable to concrete or timber structural elements as well The effectiveness of SAT is discussed by means of experimental tests on Apulian tuff specimens having known internal defects The results obtained by the new SAT approach are compared to those obtained by the ordinary amplitude tomography and by the classical travel time tomography

Fractional Fourier Transform Based Adaptive Filtering Techniques for Acoustic Emission Signal Enhancement

Abstract: Acoustic emission technique (AET) is a well-known non-destructive testing method used for the detection of crack growth and monitoring structural integrity of components. The main limitation in the application of AET is the fact that acoustic emission (AE) signal is often affected by background noise. A new approach is proposed in this paper to reduce the background noise and to improve the accuracy of AE signal detection. Various noise reduction methods have been applied and studied for AE signal enhancement so far. Some of the studies proved that self-adaptive noise cancellation (SANC) and adaptive noise cancellation (ANC) are better techniques for denoising the AE signal. This paper proposes the use of ANC and SANC schemes based on Fractional Fourier Transform (FrFT) for AE signal enhancement. FrFT is the generalization of the classical Fourier Transform (FT). The use of FrFT gives an advantage of an additional degree of freedom (angle of rotation in time–frequency plane) over the traditional Fourier transform and could provide improved performance. In this method, the noisy signal is rotated in time- frequency plane to extract the signal in Fractional Fourier domain (FrFD). Two adaptive filters viz. least mean squares and normalized least mean squares are studied for FrFD based ANC approach. The performance of the proposed method is validated using real AE signals acquired from three different experiments: pencil lead break test, composite drilling test and concrete compression test. Results are compared with the time and frequency domain adaptations. Moreover, the results have also been compared with the frequency domain adaptation techniques. Results have demonstrated that FrFD based ANC approach outperforms the conventional time domain and frequency domain adaptive filtering with an improved signal to noise ratio and significantly reduces mean square error.

Impact Damage Detection in CFRP Composite via Electrical Resistance Tomography by Means of Statistical Processing

Abstract: This study assesses the damage-detection capability of electrical resistance tomography for carbon fiber-reinforced polymer composites damaged by barely visible impact damage. Although electrical resistance tomography suffers from low spatial resolution, the method is sensitive to small conductivity changes and is therefore a suitable tool for detection of point inhomogeneity. For assessment purposes, a set of rectangular specimens were damaged by barely visible impact damage at various energies. From reconstructed images of the conductivity changes, the amplitudes and position errors for different image priors were evaluated. Amplitudes were related to the area of the delamination obtained by C-scan and assessed statistically by probability-of-detection curves. The impacts were also compared with drilled thru-holes on specimens with the same configuration. The results indicate that electrical resistance tomography is able to detect the delamination area caused by barely visible impact damage in this particular laboratory case using the amplitude of the center of gravity. The presented amplitude sensitivity, together with the position error, allows the electrical resistance tomography to be considered as a tool for structural health monitoring for early detection of barely visible impact damage what provides suggestions for further research. The performed study is the first one this comprehensive because of the number of specimens, which goes toward a statistical assessment of the electrical resistance tomography detection capability.

Comparison of Experimental Measurements of Material Grain Size Using Ultrasound

Abstract: Material grain size is related to metallic material properties and its elastic behaviour. Measuring and monitoring material grain size in material manufacturing and service is an important topic in measurement field. In this paper, three materials, i.e., aluminium 2014 T6, steel BS970 and copper EN1652, were chosen to represent materials with small, medium and large grain size, respectively. Various techniques of measuring material grain size were demonstrated and compared. These techniques include the measurements from material microstructure images, backscattered ultrasonic grain noise using a conventional transducer, longitudinal wave attenuation using ultrasonic arrays and shear wave attenuation using a lead zirconate titanate (PZT) plate. It is shown that the backscattered ultrasonic noise measurement and material attenuation measurement are complementary. The former is pretty good for weak scattering materials, e.g., aluminium, while the latter for materials with large grains, e.g., steel and copper. Consistent measured grain size from longitudinal and shear wave attenuations in steel and copper suggests that shear wave attenuation can be calculated from the measured longitudinal wave attenuation integrated with Stanke–Kino’s model or Weaver’s model, if there is a difficulty to either excite or capture shear waves in practice. The outcome of the paper expects to provide a further step towards the industrial uptake of these techniques.

A Mode-Switchable Guided Elastic Wave Transducer

Abstract: Increasing the safety demands of various progressively complex structures (e.g., in aerospace, automotive engineering, and wind power technology) and the parallel goal of reducing inspection costs set new objectives in the field of ultrasonics and guided elastic wave measurements. The focus is to deliver applicable sensors attached permanently to a particular structure for on-demand inspections. Piezoelectric fiber patches (PFPs) are known as lightweight and structure-conforming transducers for guided elastic wave generation and detection. We propose a new PFP concept combining advantages of different previous designs. With the new approach, a user can switch between a Lamb wave and a shear horizontal wave excitation by partially changing the polarity of the transducer electrodes. After summarizing the existing PFP variants, the new concept is introduced and experimentally verified by a three-dimensional laser Doppler vibrometer visualization of excited waves. Moreover, the first prototype of the new PFP transducer is presented.

Detecting Delaminations in Semitransparent Glass Fiber Composite by Using Pulsed Infrared Thermography

Abstract: Thanks to its good strength/mass ratio, a glass fibre reinforced plastic (GFRP) composite is a common material widely used in aviation, power production, automotive and other industries. In its turn, active infrared (IR) nondestructive testing (NDT) is a common inspection technique for detecting and characterizing structural defects in GFRP. Materials to be tested are typically subjected to optical heating which is supposed to occur on the material surface. However, GFRP composite is semi-transparent for optical radiation of both visual and IR spectral bands. Correspondingly, the inspection process represents a certain combination of both optical and thermal phenomena. Therefore, the known characterization algorithms based on pure heat diffusion cannot be applied to semi-transparent materials. In this study, the phenomenon of GFRP semi-transparency has been investigated numerically and experimentally in application to thermal NDT. Both Xenon flash tubes and a laser have been used for thermal stimulation of opaque and semi-transparent test objects. It has been shown that the penetration of optical heating radiation into composite reduces detectability of shallower defects, and the signal-to-noise ratio can be enhanced by applying the technique of thermographic signal reconstruction (TSR). In the inspection of the semi-transparent GFRP composite, the most efficient has been the laser heating followed by the TSR data processing. The perspectives of defect characterization of semi-transparent materials by using laser heating are discussed. A neural network has been used as a candidate tool for evaluating defect depth in composite materials, but its training should be performed in identical with testing conditions.

Local Ultrasonic Resonance Spectroscopy: A Demonstration on Plate Inspection

Abstract: Local ultrasonic resonance spectroscopy (LURS) is a new approach to ultrasound signal analysis, which was necessitated by a novel inspection method capable of the contact-free, localized, broadband generation and detection of ultrasound. By performing a LURS scan, it is possible to detect local mechanical resonances of various features and of the specimen itself. They are highly sensitive to local mechanical properties. By observing different parameters in the frequency spectrum (e.g., resonance amplitude and resonance peak frequency), geometrical, material and condition properties can be visualized for all of the scanning positions. We demonstrate LURS for inspection of a carbon fiber reinforced polymer plate. Local defect resonances of delaminations and a flat-bottom hole were detected in the frequency range from 25 to 110 kHz. Analyzing the higher frequency range (0.3 MHz to 1.5 MHz) of the same scan, the shift of the thickness resonance frequency of the plate and its higher-order resonance frequencies carry the information about the aluminum inclusions. LURS shows an advantage in characterizing the localized features of the specimens via contact-free ultrasonic inspection.

Quantitative Detection of Shallow Subsurface Defects by Using Mode-Converted Waves in Time-of-Flight Diffraction Technique

Abstract: An easy-to-implement method based on mode-converted waves was proposed for quantitatively detecting shallow subsurface defects using the ultrasonic time-of-flight diffraction (TOFD) technique. First, an isotropic model with a shallow subsurface defect was established to analyze the ray path of the mode-converted wave in TOFD B-scan image. The mode-converted wave with the shortest travel time and highest amplitude was identified to calculate the flaw depth via Snell's law. Subsequently, the quantitative expression of flaw depth was corrected by introducing the travel time of mode-converted wave, improving the measurement accuracy of the flaw height. Simulated results showed that the depth of the dead zone in carbon steel was reduced from 5.5 to 2.4 mm. The measurement error of shallow subsurface cracks was no more than 0.02 mm when the crack height was greater than or equal to 0.7 mm. Finally, three shallow subsurface defects were experimentally detected with the mode-converted waves in B-scan image. The quantitative errors of the measured heights were within 8.5% of the actual values, verifying the validity of the proposed method. This method is appropriate for the quantitative detection of shallow subsurface defects in thick-walled components.

Scale–Space Data Augmentation for Deep Transfer Learning of Crack Damage from Small Sized Datasets

Abstract: Crack damage is commonly observed for civil structures and infrastructure in service. The recent years have witnessed an excessive utilization of deep learning models for realizing autonomous and machine-vision based crack detection. Given this trend, this paper recognizes two entwined challenges: the preparation of large-scale training data and the detection of simple crack damage amid complex scenes. To address them, a novel data augmentation technique is proposed considering crack characteristics in images for realizing deep transfer learning using very small datasets. Numerical experimentation is conducted based on two types of crack datasets (concrete structures and asphalt pavement), each of which has only tens of images containing complex scenes. When evaluating the performance, a sliding-window based rating scheme is proposed, which is much stricter than the conventional bounding-box based approach. Quantitative performance analysis shows the acceptable performance (e.g., an overall accuracy of 93.81%, an F-2 score of 74.4%, and a very high recall of 91% for the crack detection in concrete images). The result demonstrates the effectiveness of the proposed data augmentation method and the superior transferability if the transfer learning is carried out through a fully fine-tuned training process.

A Novel Magnetic Flux Leakage Testing Method Based on AC and DC Composite Magnetization

Abstract: The flexible printed coil array (FPC) is employed extensively because of its flexibility in non-destructive testing (NDT). To make the inductive coil detect both external and internal defects under static conditions in magnetic flux leakage testing (MFL), this paper proposes a method based on the AC and DC composite magnetization. The AC magnetization changes the magnetic field strength in the vicinity of external defects. The DC magnetization makes the defects generate static leakage magnetic field. The permeability in the skin layer is also altered by the AC magnetization which affects the transmission of leakage magnetic field caused by internal defects. Therefore, external and internal defects can generate a time-varying leakage magnetic field. Simulations and experiments are carried out to validate the feasibility of this method. The results indicate that the method based on the AC and DC composite magnetization can be applied to detect both external and internal defects.

Use of Bacteria as Fluorescent Penetrant for Penetrant Testing (PT)

Abstract: Penetrant testing (PT) is a Nondestructive testing (NDT) method used for identifying and revealing surface cracks and defects in components. Using bacteria-based suspension as a fluorescent penetrant leads a new approach to PT with the advantages of using eco-friendly chemicals, fewer steps for test process and achieving low cost. The aim of the current study is to synthesize a new fluorescein isothiocyanate (FITC) conjugated Escherichia coli (E. coli) bacteria (FITC-E. coli) as fluorescent penetrant for detection of cracks on test materials. For this purpose, we tested performance of proposed bacteria suspension by means of its detection sensitivity using standardize test panels. Crack centers (6.35, 3.97 and 2.38 mm) and defects (0.2 × 0.4, 0.3 × 0.4 and 0.4 × 0.4 mm) were successfully detected via FITC-E. coli. Experimental results showed that the FITC-E. coli is efficient for detecting the cracks and defects. Research into development of bacteria-based fluorescent penetrant may bring new opportunities for NDT applications in the future.

Distributed Fiber Optics Sensing Applied to Laminated Composites: Embedding Process, Strain Field Monitoring with OBR and Fracture Mechanisms

Abstract: Optical fibers (OFs) are among the most promising technologies for distributed sensing in structure health monitoring of composites. Optical backscatter reflectometer (OBR) based on Rayleigh scattering enable to use the entire length of telcom OF as sensor, giving detailed information about the stain fields in the host structure. Embedded sensors can provide many advantages over surface bonded ones in terms of reliability and accuracy of the measurements. Here, OFs were embedded in between (0/0), (0/90) and (90/90) interfaces of a [02/902/02] glass fiber reinforced polymer (GFRP) and were further interrogated with OBR during tensile tests. Sensors were evaluated in terms of its sensing capabilities and fracture mechanism. OFs aligned in between (0/0) direction endured 5500 µe and deviated only 2.7% in the Young’s modulus calculation compared to mounted strain gauges. The other directions presented problems related to matrix rich area (0/90) and air bubbles imprisoning (90/90), reducing the sensing range. After reaching the cracking state, OFs fail progressively and are still capable of acting as sensor, as was showed in a simple visible laser scattering test. For the first time, embedding, interrogation with OBR and fracture mechanisms are approached in a single study, which provide valuable information for wide application of OFs in SHM.

High-Speed Acoustic Impact-Echo Sounding of Concrete Bridge Decks

Abstract: Evaluation of deteriorating highway bridges requires new, rapid inspection methods to effectively guide repair in an era of limited fiscal resources. Of all the components of a bridge, the bridge deck typically deteriorates most quickly and must therefore be regularly inspected for non-visible internal cracking, called delamination, for which early detection and repair can enhance safety and performance as well as reduce long-term maintenance costs. This research demonstrates the development and deployment of a multi-channel acoustic impact-echo sounding apparatus equipped with specially-configured tire chains for acoustic excitation of the concrete, microphones, an array of sensors for establishing spatial position estimates, and a data processing architecture for accurately detecting and mapping delaminations in concrete bridge decks at high speeds. The new apparatus achieved very accurate results at speeds between 25 and 45 km/h across a bridge deck. These results demonstrate an orders-of-magnitude increase in data collection speed over all other acoustic impact-echo sounding techniques described in the literature.

Concrete Compressive Strength Prediction Using Neural Networks Based on Non-destructive Tests and a Self-calibrated Response Surface Methodology

Abstract: An artificial neural network (ANN) model and response surface methodology (RSM) were established to estimate the compressive strength of concrete by using the combination of three non-destructive tests (NDT); rebound number, pulse velocity tests and resistance surface. These techniques are utilized in an attempt to increase the reliability of the non-destructive tests in detecting the strength of concrete. These methods were trained using a set of different mixes and at different ages of concrete specimens. In this case, 180 experimental specimens were conducted and their data are published. Then, different neural network topologies and algorithms besides RSM were examined using the given data. The published models are for two combination including the combination of UPV and RN and the combination UPV, RN and SR. The results show that the accuracy of the published models are increased by aging. In addition, it is showed that RSM don’t need calibration process, while its accuracy is enough. Hence, RSM is a promising method to conduct on NDTs and compressive strength prediction, while ANN needs to perform many times to find the best accuracy.

Microwave Non-destructive Evaluation of Glass Reinforced Epoxy and High Density Polyethylene Pipes

Abstract: Owing to their inherent advantages, non-metallic structures such as glass reinforced epoxy (GRE) pipes and high-density polyethylene (HDPE) pipes are rapidly replacing their steel counterparts in pipeline infrastructures. However, non-metallic pipes are vulnerable to a variety of defects which compromise their structural integrity. Hence, periodic inspection of these structures is incumbent to ensure their structural integrity and avoid catastrophic in-service failures. Microwave and millimeter wave non-destructive evaluation techniques have shown great potential for inspecting non-metallic structures due to their high sensitivity, resolution, and cost-effectiveness. However, most of the previous investigations performed on microwave non-destructive testing (NDT) of these particular structures are deemed limited since these investigations considered one probe type and often conducted over narrow frequency band. Furthermore, many of the previous microwave NDT investigations considered unrealistically large defects which renders the reported conclusions about the performance envelope of microwave NDT of non-metallic pipes rather inaccurate. Consequently, an uncertainty still exists in the domain with regards to detection capability of the method when the defects in the pipes are actually small in size. Unlike all previous investigations, a comprehensive and realistic microwave imaging of GRE and HDPE pipes with embedded defects of practical importance is demonstrated in this paper. In order to image these defects, various near-field probes operating at wide range of frequencies are employed and compared herein. The relative performance and the effectiveness of microwave NDT for non-metallic pipe inspection using these probes are quantitatively reported. Furthermore, the images of the HDPE pipe produced by microwave probes are benchmarked to phased array ultrasonic testing.

Wavelet Based Clustering of Acoustic Emission Hits to Characterize Damage Mechanisms in Composites

Abstract: Clustering data is an important topic in acoustic emission (AE) health monitoring. Identification of damage mechanisms in composites via AE technique requires an automatic process of classification. In this work, we propose a novel and simple supervised method to discriminate AE signals produced by fracture mechanisms in polymer composites. The novelty of this work is to propose new pertinent descriptors offered by using the continuous wavelet transform, where signals of learning are decomposed and the corresponding wavelet coefficients are calculated. In addition, the entropy criterion is applied to select the most correlated wavelets associated to each failure mechanism. This process allows to establish a filter in the form of vectors for each class of signals and descriptors denote the reconstruction errors calculated by involving the filter associated to each damage mechanism. The k-means algorithm is executed to calculate the center of each class. The technique is applied to AE signals recorded from specific mechanical tests to demonstrate the performance of the proposed descriptors.

Operational Damage Identification Scheme Utilizing De-Noised Frequency Response Functions and Artificial Neural Network

Abstract: A damage identification scheme combining impact-synchronous modal analysis (ISMA) and artificial neural network is developed in this study. The ISMA de-noising method makes it feasible to detect and classify the damage states with high accuracy when the machine is under operation. The feed-forward backprop network was utilized in this study. The input feature vector of the network consisted of the FRF changes in a selected vibrational mode frequency interval at several measurement points. The scheme was tested on a rectangular Perspex plate. It is proved that the trained network can successfully identify damage locations with the testing data collected by ISMA, which allows the damage detection to be carried out without shutting down the tested machine. For the plate structure in this study, an overall accuracy reached 100% when all five measurement points were used. With the input features optimized by mode shape assessment, 100% accuracy was also achieved with only two measurement points.

Pulsed Eddy Current: Feature Extraction Enabling In-situ Calibration and Improved Estimation for Ferromagnetic Application

Abstract: Steel pipes in process plant applications are often covered with insulation or weather protection that make inspection difficult because the additional layers need to be penetrated to inspect the pipes’ structure. The pulsed eddy current (PEC) method was devised as a means of inspection through the surface layers. However, the performance of a PEC system is dependent on the electrical and magnetic properties of the pipe material, which are generally unknown. Therefore, the use of a calibration block from a different steel will give inaccurate results. The concept of calibrating using $$\tau _0$$ values obtained during inspection has undoubtedly been discussed in the literature. However, no comprehensive work was dedicated to using $$| abla |^{-1}$$ to carry out calibration on inspected structure. The linear relationship of the $$| abla |^{-1}$$ feature with the thickness squared, $$d^2$$ , is first established using analytical solutions, and the calibration is carried out using the feature values obtained in air and the reference signal. The performance of this technique is assessed and compared with the conventional $$\tau _0$$ technique. Although both features exhibit similar immunity towards lift-off, $$\tau _0$$ technique requires normalisation procedure, which contributes to determining more configuration parameters. Experimental results also suggest the relative advantage of using $$| abla |^{-1}$$ feature in both wall thickness estimation and influences of noises.

Ultrasonic Nondestructive Evaluation of Porosity Size and Location of Spot Welding Based on Wavelet Packet Analysis

Abstract: There is a lack of efficient methods for ultrasonic nondestructive quantitative evaluation of porosity in resistance spot welding, and the precise location of porosity in a nugget can be determined only when the sizes of porosity and nugget are calculated concurrently. The sizes of both porosity and nugget in a resistance spot welded joint made from stainless steel based on the decomposition and reconstruction technique of wavelet packet were investigated. The newly reconstructed sequences in some decomposed nodes make the porosity more salient than the original signal. The signals in different porosities reflected by a fixed wavelet packet node were found to be inadequate. Thus, the wavelet packet coefficients in different nodes were revised to reconstruct a new signal. This method is less affected by the conditions of workpiece surface than the original signal and application of a fast Fourier transform. The porosity size and location can be calculated more accurately, and the calculation precision is limited to 7%.

Intelligent Condition Monitoring of Ball Bearings Faults by Combination of Genetic Algorithm and Support Vector Machines

Abstract: Bearings are one of the most widely used components in the industry that are more vulnerable than other parts of machines. In this research, a precise method was developed for diagnosis bearing detection based on vibrating signals. Vibration signals were recorded from four common faults in the bearings at three speeds of 1800, 3900, and 6600 rpm. The vibration signals were transmitted by the fast Fourier transform to the frequency domain. A total of 24 features were extracted from frequency and time signals. The superior features are selected using the combination of genetic algorithm and artificial neural network. A support vector machine is used to intelligently detect ball bearing faults. The accuracy of the support vector machine with all extracted features in different revolutions showed that the highest accuracy for training and test data was obtained 78.86% and 69.33% respectively, at 1800 rpm. The results of reduction and selection of superior features showed that the highest accuracy of the support machine was obtained in the classification of ball bearing faults for training and test data 97.14% and 93.33%, respectively. The results show that the use of the feature selection method based on the genetic algorithm will increase the accuracy of the classification.