Surface-breaking Defects Research Articles

Shear Wave Ultrasound Inspection of Flaws in Silicon Wafers using Focused Transducers.

Silicon parts can contain micrometer-sized vertical cracks that are challenging to detect. Inspection using high-frequency focused ultrasound has shown promise for detecting defects of this size and geometry. However, implementing focused ultrasound to inspect anisotropic media can prove challenging, given the directional dependence of wave propagation and subsequent focusing behavior. In this work, back surface-breaking defects at various orientations within silicon wafers (0-, 15-, and 45-degrees relative to the [010] crystallographic axis) are experimentally inspected in an immersion tank setup. Using 100 MHz unfocused and focused shear waves, the impact of medium anisotropy on focusing and defect detection is evaluated. The scattering amplitude and defect detection sensitivity results demonstrate orientation-dependent patterns that strongly rely on the use of focused transducers. The defects along the 45-degree orientation reveal two-lobe scattering patterns with maximum amplitudes less than half that of the defects in the 0-degree orientation, which in contrast show a one-lobe scattering pattern. The experimental results are further explored using Finite Element (FE) modeling and ray tracing to visualize the impact of focusing on wave propagation within the silicon. Ray tracing results show that the focused beam profiles for the 45- and 0-degree orientations form a butterfly wing and elliptical focusing profile, respectively, which correspond directly to experimentally found scattering patterns from defects. Additionally, the FE scattering results from unfocused transducers reveal single lobe scattering for both 0- and 45-degree orientations, proving the varying scattering patterns to be driven by the anisotropic focusing behavior.

How to Reach 100% Inspection Coverage of Aeroengine Fan Blades with a High Probability of Detection

In aviation, quality is paramount to determine an aircraft’s airworthiness and keep personnel and passengers safe. Non-destructive testing (NDT) methods are largely used throughout an aircraft’s lifetime, including raw material selection, manufacturing, assembly, and in-service maintenance. As a key component of aircraft, engines are governed under strictly scheduled maintenance programs where they are dismantled and thoroughly inspected. Part of this global inspection concerns the individual verification of fan blades to assess the absence of flaws, such as surface cracks. Due to the complex geometry of fan blades, established methods such as penetrant testing (PT) and magnetic particle inspection (MPI) have been successfully used in past decades to detect surface-breaking flaws in fan blades. However, these techniques are complex to deploy and do not permit results traceability. As an alternative, eddy current (EC) technologies proved to be a good candidate to discriminate flaws as small as 1 mm long. Combining both conventional and eddy current array technology provides valuable information on the integrity of the component. Powered with the development of dedicated signal processing methods, the proposed solution helped to increase signal-tonoise ratio, avoiding any false calls. Once this was achieved, a careful study of the probe’s deployment and compliance with the various changes of geometries—from the blade’s surface to its edges and root, helped ensure optimal crack detection. Finally, recent advances in robotics enabled an automated inspection strategy to perform 100% coverage with high reproducibility and repeatability. This greatly improved the probability of detection of sub-millimeter surface-breaking cracks. This paper describes the study performed to assess the capability of eddy current solutions (conventional and advanced) to replace traditional methods of inspecting titanium fan blades. It also shows how dedicated signal processing enables increased probability of detection. The proposed approach to provide full coverage of the complex geometry and the associated developments are detailed. Finally, the paper presents an integration of the complete inspection technique in an automated system.

Numerical Study of Rayleigh Wave Interaction with Rolling Contact Fatigue Type of Defects

ABSTRACT Rolling Contact Fatigue or Damage (RCF/RCD) presents significant maintenance challenges to railroads across the globe. Quantifying RCF/RCD crack depths and density in rails is important for all railroads to manage their grinding programs effectively and efficiently and being able to conduct ultrasonic testing (UT) of rails for reliable detection of internal fatigue damage. This work focuses on the modeling of Rayleigh waves UT approach to detect and characterize RCF type of defects which can form as: vertical, oblique or branched shaped surface breaking defects in the rail head. Specifically, the transmission coefficient (Tc) of Rayleigh waves was studied using finite element analysis (FEA). The effect of crack tip geometry on Tc values is discussed. The results suggest that for oblique, and branch cracks, characterization based purely on the Tc can be challenging due to symmetric sinusoidal fluctuations in the Tc. A real crack using a micrograph image was also modeled to validate the Tc results for oblique and branched cracks. This points to the need for additional parameters to be identified for efficient and reliable characterization of RCF/RCD type of defects in rails.

Tolerance of oxide wedging crack growth of manufacturing flaws in CANDU Zr-2.5Nb pressure tubes to cyclic operating stresses

In-service inspections of Zr-2.5Nb pressure tubes in CANDU nuclear reactors have detected inner diameter surface breaking flaw indications which are classified as manufacturing flaws. It is considered prudent to evaluate potential growth by means of a postulated oxide wedging crack growth mechanism. An oxide wedging crack growth model has previously been developed and applied to predict future flaw sizes. Fatigue crack growth due to cyclic operating stresses is calculated for these flaws, and this is added to the postulated oxide wedging crack growth for fitness-for-service evaluations. This established approach is overly conservative since fatigue crack growth cannot occur while the oxide wedging crack growth mechanism is active, due to the essentially displacement-controlled loading of the flaw by the oxide wedge, which causes the stress intensity factor for the flaw to be effectively constant under cyclic operating stresses. This paper addresses this conservatism by evaluating crack growth due to postulated oxide wedging, and demonstrating that fatigue crack growth is not possible under oxide wedge loading.

Analysis of the Effect of Machining of the Surfaces of WAAM 18Ni 250 Maraging Steel Specimens on Their Durability.

It is now well-known that the interaction between surface roughness and surface-breaking defects can significantly degrade the fatigue life of additively manufactured (AM) parts. This is also aptly illustrated in the author's recent study on the durability of wire and arc additively manufactured (WAAM) 18Ni 250 Maraging steel specimens, where it was reported that failure occurred due to fatigue crack growth that arose due to the interaction between the surface roughness and surface-breaking material defects. To improve the durability of an AM part, several papers have suggested the machining of rough surfaces. However, for complex geometries the fully machining of the entire rough surface is not always possible and the effect of the partial machining on durability is unknown. Therefore, this paper investigates if partial machining of WAAM 18Ni 250 Maraging steel surfaces will help to improve the durability of these specimens. Unfortunately, the result of this investigation has shown that partial machining may not significantly improve durability of WAAM 18Ni 250 Maraging steel specimens. Due to the order of surface roughness seen in WAAM 250 Maraging steel, the improvement to durability is only realized by full machining to completely remove the remnants of any print artefacts.

Dynamic Inspection of Surface-Breaking Defects using Induction Thermography

In comparison to other active thermography techniques, induction thermography can provide instantaneous results, is not affected by material emissivity/reflectivity and is indifferent to surface geometry. These advantages make it a promising technique to be applied for in-line/dynamic inspections. In this work, the thermal signatures from a titanium sample under different excitation parameters and moving speeds will be studied. To simulate surface-breaking cracks, electrical discharge machining (EDM) notches of varying lengths, widths and depths were made on the titanium sample. As with most applications, only a single side is accessible for inspection. Therefore, induction thermography in the reflection mode will be utilized in this work. After evaluating the signal-to-noise ratio (SNR) under various conditions and inspection speeds, the optimized configuration will be presented. The findings will be beneficial to researchers interested in implementing induction thermography in an in-line/dynamic inspection setting.

Automated Real-Time Eddy Current Array Inspection of Nuclear Assets.

Inspection of components with surface discontinuities is an area that volumetric Non-Destructive Testing (NDT) methods, such as ultrasonic and radiographic, struggle in detection and characterisation. This coupled with the industrial desire to detect surface-breaking defects of components at the point of manufacture and/or maintenance, to increase design lifetime and further embed sustainability in their business models, is driving the increased adoption of Eddy Current Testing (ECT). Moreover, as businesses move toward Industry 4.0, demand for robotic delivery of NDT has grown. In this work, the authors present the novel implementation and use of a flexible robotic cell to deliver an eddy current array to inspect stress corrosion cracking on a nuclear canister made from 1.4404 stainless steel. Three 180-degree scans at different heights on one side of the canister were performed, and the acquired impedance data were vertically stitched together to show the full extent of the cracking. Axial and transversal datasets, corresponding to the transmit/receive coil configurations of the array elements, were simultaneously acquired at transmission frequencies 250, 300, 400, and 450 kHz and allowed for the generation of several impedance C-scan images. The variation in the lift-off of the eddy current array was innovatively minimised through the use of a force–torque sensor, a padded flexible ECT array and a PI control system. Through the use of bespoke software, the impedance data were logged in real-time (7 ms), displayed to the user, saved to a binary file, and flexibly post-processed via phase-rotation and mixing of the impedance data of different frequency and coil configuration channels. Phase rotation alone demonstrated an average increase in Signal to Noise Ratio (SNR) of 4.53 decibels across all datasets acquired, while a selective sum and average mixing technique was shown to increase the SNR by an average of 1.19 decibels. The results show how robotic delivery of eddy current arrays, and innovative post-processing, can allow for repeatable and flexible surface inspection, suitable for the challenges faced in many quality-focused industries.

Uncertainty Quantification for Deep Learning in Ultrasonic Crack Characterization.

Deep learning for nondestructive evaluation (NDE) has received a lot of attention in recent years for its potential ability to provide human level data analysis. However, little research into quantifying the uncertainty of its predictions has been done. Uncertainty quantification (UQ) is essential for qualifying NDE inspections and building trust in their predictions. Therefore, this article aims to demonstrate how UQ can best be achieved for deep learning in the context of crack sizing for inline pipe inspection. A convolutional neural network architecture is used to size surface breaking defects from plane wave imaging (PWI) images with two modern UQ methods: deep ensembles and Monte Carlo dropout. The network is trained using PWI images of surface breaking defects simulated with a hybrid finite element / ray-based model. Successful UQ is judged by calibration and anomaly detection, which refer to whether in-domain model error is proportional to uncertainty and if out of training domain data is assigned high uncertainty. Calibration is tested using simulated and experimental images of surface breaking cracks, while anomaly detection is tested using experimental side-drilled holes and simulated embedded cracks. Monte Carlo dropout demonstrates poor uncertainty quantification with little separation between in and out-of-distribution data and a weak linear fit ( R=0.84 ) between experimental root-mean-square-error and uncertainty. Deep ensembles improve upon Monte Carlo dropout in both calibration ( R=0.95 ) and anomaly detection. Adding spectral normalization and residual connections to deep ensembles slightly improves calibration ( R=0.98 ) and significantly improves the reliability of assigning high uncertainty to out-of-distribution samples.

A Study On Effects of Flaw Shape Idealization On the Interaction of Co-Planar Surface Flaws Subjected to Tension Load

Abstract Engineering Critical Assessment (ECA) guidelines contain rules to assess flaw interaction. Major flaw dimensions (depth or height and length) are typically characterized assuming the flaws to be contained entirely within a bounding rectangle through a procedure known as flaw idealization. In fracture mechanics based calculations, flaws are often assumed to be (semi-)elliptical when evaluating possible interaction. This paper investigates the implication of this simplification for the specific case of two identical co-planar surface breaking flaws. Two flaw shapes are considered and compared: semi-elliptical and canoe-shaped (quarter-circular ends with constant depth elsewhere). Especially for long and shallow flaws, the canoe-shaped configuration best approximates the bounding rectangle, whereas the semi-elliptical shape only touches the bounding rectangle at three points (deepest point and two points at the surface). Several flaw dimensions and spacing distances are studied through an extensive parametric study comprising both linear elastic and elastic-plastic finite element simulations. The results, evaluated in terms of stress intensity factor (SIF) and J-integral, show that the flaw shape idealization, particularly for long and shallow flaws, can significantly affect the degree of interaction between identical co-planar flaws. The inconsistency between semi-elliptical and canoe-shaped flaw shapes is observed in a linear elastic analysis and becomes more pronounced at higher loading levels evaluated in elastic-plastic analyses

A novel laser-based duffing oscillator system to identify weak ultrasonic guided wave signals related to rail defects

Laser-generated Rayleigh wave is a promising candidate for noncontact and remote inspection of railway as it can inspect in line-to-line fashion. Attribute to the recent advancement of optical design, wavelength of the laser generated-Rayleigh wave can be controlled at desired frequency band. By adjusting the wavelength of the Rayleigh wave, detection of defect occurred at different depths in rail become possible. Internal and subsurface defects that occur beneath the surface of railhead usually only reflect very small portion of wave energy hence accurate determination of their presences becomes challenging. Besides, such reflected-defect signal is usually contaminated with strong random noise which further hinder identification of weak reflected-defect signal. This study presents a novel method for detection of surface, subsurface and internal defects in railhead using laser-generated Rayleigh wave and chaotic oscillator. The use of laser-generated Rayleigh wave provides a fast mean to assess the health condition of railhead without the need to physically mounting sensor on the specimen. The use of chaotic oscillator can accurately reveal wave signal that is related to defect occurred in railhead. The effectiveness of the proposed method has been verified using numerical simulation model. Three rail specimens were machined with internal, subsurface and surface-breaking defects, respectively. Experimental results revealed that the proposed method can accurately and reliably reveal the presence of reflected-defect signal under strong random noise conditions. The present study also shows that the proposed system is promising in revealing hazardous defects occurred on rai

Domain Adapted Deep-Learning for Improved Ultrasonic Crack Characterization Using Limited Experimental Data.

Deep learning is an effective method for ultrasonic crack characterization due to its high level of automation and accuracy. Simulating the training set has been shown to be an effective method of circumventing the lack of experimental data common to nondestructive evaluation (NDE) applications. However, a simulation can neither be completely accurate nor capture all variability present in the real inspection. This means that the experimental and simulated data will be from different (but related) distributions, leading to inaccuracy when a deep learning algorithm trained on simulated data is applied to experimental measurements. This article aims to tackle this problem through the use of domain adaptation (DA). A convolutional neural network (CNN) is used to predict the depth of surface-breaking defects, with in-line pipe inspection as the targeted application. Three DA methods across varying sizes of experimental training data are compared to two non-DA methods as a baseline. The performance of the methods tested is evaluated by sizing 15 experimental notches of length (1-5 mm) and inclined at angles of up to 20° from the vertical. Experimental training sets are formed with between 1 and 15 notches. Of the DA methods investigated, an adversarial approach is found to be the most effective way to use the limited experimental training data. With this method, and only three notches, the resulting network gives a root-mean-square error (RMSE) in sizing of 0.5 ± 0.037 mm, whereas with only experimental data the RMSE is 1.5 ± 0.13 mm and with only simulated data it is 0.64 ± 0.044 mm.

Analysis on asymmetrical RCF cracks characterisation using an ACFM sensor and the influence of the rail head profile

Rolling contact fatigue (RCF) cracks are one of the typical surface-breaking defects that impairs the rail service life. RCF cracks propagation can deviate from regular semi-ellipse shapes, turning into asymmetrical profiles. RCF crack measurements using the alternating current field measurement (ACFM) method are unable to accurately characterise asymmetrical cracks, as the sizing algorithm is developed based on regular crack shapes. In this paper, the development of modelling capability to consider asymmetrical RCF cracks and the influence of asymmetrical shapes on ACFM signals are presented. A novel grid scan method using ACFM probe sensor for characterizing asymmetrical crack shapes is proposed. The influences of rail head profile on crack sizing are studied and result shows that it gives maximum 2 mm underestimation in crack depth if the worn rail head profile is not considered in ACFM sizing algorithm.

Exploring high-frequency eddy-current testing for sub-aperture defect characterisation using parametric-manifold mapping

Accurate characterisation of small defects remains a challenge in non-destructive testing (NDT). In this paper, a principle-component parametric-manifold mapping approach is applied to single-frequency eddy-current defect characterisation problems for surface breaking defects in a planar half-space. A broad 1–8 MHz frequency-range FE-circuit model & calibration approach is developed & validated to simulate eddy-current scans of surface-breaking notch defects. This model is used to generate parametric defect databases for surface breaking defects in an aluminium planar half-space and defect characterisation of experimental measurements performed. Parametric-manifold mapping was conducted in N-dimensional principle component space, reducing the dimensionality of the characterisation problem. In a study characterising slot depth, the model & characterisation approach is shown to accurately invert the depth with comparable accuracy to an equivalent peak amplitude-phase inversion method. The approach is used to characterise the depth of a sloped slot demonstrating good accuracy up to ≈ 2.0 mm in depth over a broad range of frequencies, indicating applications in geometric feature inversion. Finally the technique is applied to finite rectangular notch defects of surface extents smaller than the diameter of the inspection coil (sub-aperture) over a range of sub-resonant frequencies. The results highlight the limitations in characterising these defects and indicate how the inherent modelling uncertainty around resonance can severely limit characterisation at these frequencies.

Optimising sensor pitch for magnetic flux leakage imaging systems

Sensor arrays can significantly increase the speed at which inspections and subsequent imaging of flaws is performed[1]. This work focuses on developing a software approach for optimising the spacing between quantum well Hall-effect (QWHE) magnetic sensors used for magnetic flux leakage (MFL) imaging, where this approach could be adapted for any non-destructive evaluation (NDE) technique in which imaging is obtained. A ground mild steel weld sample containing two surface-breaking flaws prepared by Sonaspection was scanned using an XYZ MFL imaging system developed at the University of Manchester[2,3,13,14]. The scan was taken with an autonomously controlled lift-off height of 0.75 mm, with an x-y measurement step of 100 μm and an applied magnetic field of 30 mT root mean square (RMS) at a frequency of 400 Hz. This data (ie magnetic image) was then processed to simulate different measurement step sizes, to determine any relationship between step size and flaw detectability (flaw signal to weld background response). This work effectively simulates different sensor pitches (separation between sensors) of integrated QWHE sensor arrays from 100 μm to 5 mm, with the goal of determining both the minimum number of sensors required in the array and the optimal spacing to maximise scan speeds and help determine optimum inspection parameters to develop the technology of low-power MFL imaging. This optimisation process could be applied to any NDE imaging system (electromagnetic or other) currently used, with results dependent on the inspection parameters.

Magnetic Flux Leakage Course of Inner Defects and Its Detectable Depth

As a promising non-destructive testing (NDT) method, magnetic flux leakage (MFL) testing has been widely used for steel structure inspection. However, MFL testing still faces a great challenge to detect inner defects. Existing MFL course researches mainly focus on surface-breaking defects while that of inner defects is overlooked. In the paper, MFL course of inner defects is investigated by building magnetic circuit models, performing numerical simulations, and conducting MFL experiments. It is found that the near-surface wall has an enhancing effect on the MFL course due to higher permeability of steel than that of air. Further, a high-sensitivity MFL testing method consisting of Helmholtz coil magnetization and induction coil with a high permeability core is proposed to increase the detectable depth of inner defects. Experimental results show that inner defects with buried depth up to 80.0 mm can be detected, suggesting that the proposed MFL method has the potential to detect deeply-buried defects and has a promising future in the field of NDT.

All-optical laser-ultrasonic technology for width and depth gauging of rectangular surface-breaking defects.

In this paper, the width and depth of rectangular surface-breaking defects were successfully gauged using an all-optical laser-ultrasonic system. The finite element method was used to simulate propagating paths of defect-induced reflected and transmitted Rayleigh waves. It is observed that most Rayleigh waves with wavelengths less than the depth of the defect are reflected by the defect. A small part of the Rayleigh waves with wavelengths greater than the depth of the defect directly transmits through the bottom edge of the defect without acting on the left edge of the defect. Based on the simulation results, a three-step detection method of width and depth gauging of rectangular surface-breaking defects is proposed. In the first step, the pulsed laser and detection laser are irradiated on one side of the defect. In the second step, the sample is moved to a certain distance to ensure that both lasers reach the other side of the defect. In the third step, two lasers are irradiated on both sides of the defect. The width and depth of the defect are calculated according to the arrival time of the incident, reflected, and transmitted Rayleigh waves, as well as the movement distance of the sample. Experimental results are consistent with the reference-results measured by using a digital microscope. The proposed three-step detection method is proved to be feasible in simultaneous measurement of the width and depth of rectangular surface-breaking defects. Furthermore, it may be potentially useful for measuring other types of defect structures.

Depth feature extraction of surface-breaking defects using laser pulsed thermography

In this paper, the feasibility of the depth feature extraction of surface-breaking defects based on laser pulsed thermography in transmission mode is proposed. This method is adopted to detect the depth of surface-breaking defects. First, a finite element model based on COMSOL is established to simulate structural steel specimens with different defect depths and the contrast method is used to detect these depths. In order to verify its feasibility, the model is simulated at different time nodes. A simulation analysis shows the practicability of this method. Then, through an artificial slot depth feature extraction experiment on structural steel specimens, an algorithm is used to denoise an infrared image and then the contrast method is used to extract the defect depth feature, which verifies the feasibility of the method. The experimental results show that as the depth of artificial slots increases, the temperature at the observation point also reduces. The best fitting equation of the defect depth and the temperature at the observation point at a certain heating time have an exponential relationship. This method can accurately detect defect depths.

Development and optimisation of low-power magnetic flux leakage inspection parameters for mild steel welds

Magnetic particle and other magnetic flux leakage (MFL)-based methods for the detection and evaluation of surfacebreaking flaws in ferromagnetic materials typically use high-strength (≥0.5 T RMS), low-frequency (≤50 Hz) magnetic fields. The rationale behind this is the ready availability of strong permanent magnets and mains power for high-strength electromagnets. This high field strength is needed to saturate the sample and compensate for the relatively low sensitivity of magnetic particle detection media, silicon Hall sensors, coils and other magnetic transducers used in such methods. Consequently, frequencies greater than 50 Hz and applied magnetic fields less than 100 mT strength have not been widely explored for MFL due to the lack of commercially available sensors capable of detecting the leakage fields (typically in the nT and μT range) with adequate versatility to cope with the variations in inspection parameters, such as changes in liftoff, material properties, etc, which are inherent to non-destructive testing and evaluation (NDT&E) settings. In this study, the MFL response of surface-breaking longitudinal cracks from a ground mild steel weld validation sample, within the DC to 1 kHz and 5 mT to 100 mT RMS applied magnetic field operating range, was explored. This was carried out to determine whether any optimal frequency response exists, better accommodating the inherent sample material properties (for example magnetic permeability and electrical conductivity) and MFL mechanism and attributing phenomena such as electromagnetic skin effect and eddy current contributions. Contrary to previous work published in Insight last year, this study found no particular optimal frequency within this operating range, with explanations to justify the disproval of previously reported conclusions about optimal frequencies within this range. Also, the iteratively developed quantitative analysis performed in this study can be used to help further understand the underlying mechanisms of AC MFL and provide best practice regarding the optimisation of MFL.

Failure analysis of a ruptured compressor pressure vessel

This paper deals with failure analysis of a ruptured pressure vessel that served as a part of an air compressor. Pressure vessel failed during hydrostatic test when, at a pressure level lower than proof pressure, two cracks occurred on the shell and test fluid leakage was recorded. Experimental investigation methods were employed to determine possible causes of crack occurrence and vessel rupture. Liquid penetrate inspection was performed to check for other surface-breaking defects on the shell and locations of the cracks. Visual examination revealed the condition of the internal surface of the shell. Optical microscopy was used to inspect crack area while scanning electron microscopy examination at suitable magnification was employed to characterize fine microstructure of the fractured surface and to reveal flaws that served as rupture initiation points. Material’s chemical composition was determined using optical emission spectrometer with glow discharge source sample stimulation. Hardness test results were used to derive maximum tensile strength of the material. According to results obtained from performed investigations, it was concluded that excessive corrosion at the bottom part of the pressure vessel caused formation of pits that served as initiation points for leakage cracks. Recommendations for avoidance of such scenario with the rest of the user’s compressor pressure vessel were given. Also, ultrasonic testing of shell thickness was performed and results of the performed hydrostatic pressure test were recorded to be used in subsequent numerical failure analysis of the failed pressure vessel.

Detecting surface-breaking flaws with a Hall effect gradiometric sensor

Magnetic scanning devices for structures, components, and materials have been the subject of intense scientific research and are potential prototypes for use in industry, especially in non-destructive inspection, to identify defects in metal structures subject to extreme conditions such as high temperatures, pressure, and strong tension. These conditions can lead to the failure of these structures, affecting their operation and resulting in losses and possible accidents. In this sense, we built a device for scanning magnetic response in defective steel plates. The device is composed of two Hall effect sensors (Melexis MLX 90215) in a gradiometric configuration. Using this device, it was possible to identify defects produced by electric discharge machining in SAE 1020 steel plates and proved to be efficient in detecting defects with diameters and depths of the order of 0.90 mm and 100 µm, respectively. A filter system on a printed circuit board was also developed to attenuate noise. The signal-to-noise analysis showed the circuit was effective, leading to identification of defects. To reinforce the application of this equipment in scientific research and industry, we developed a theoretical model to estimate the depth of defects.