Rayleigh waves can be used for characterisation of surface-breaking defects, giving a measure of the depth and the angle of propagation of a defect with simple (i.e. single crack) geometry. However, surface breaking defects will often grow with a more complicated geometry. We present here results of experimental measurements using laser generated and detected Rayleigh waves on aluminium samples containing machined slots with varied branched geometries. The signal enhancement found in the near-field, and the reflection and transmission of different wavemodes can be used to position the defect and gain an idea of its geometry. This research can be applied to monitor components prone to developing stress corrosion cracking (branched-like defects). Results are shown of the near-field interactions of Rayleigh waves with this type of cracking in stainless steel pipe samples, in order to resolve the spatial extent and geometric alignment of those defects.

The enhancement of surface waves in the near-field of a defect has been reported by several authors. It has been demonstrated that the interaction between the incident Rayleigh wave with the reflected Rayleigh wave, plus the mode-converted surface skimming longitudinal wave, explains the significant increase in signal amplitude that is encountered as a detection point approaches a smooth machined defect inclined normal to the surface. However, this is not a typical defect geometry. For example, rolling contact fatigue in rails grows at an angle to the surface, and stress corrosion cracking grows as branched defects. We present results of experimental measurements on machined slots with varied geometries, including defects which are normal or inclined to the surface, and show the effect of branched defect geometries on the wave propagation and signal enhancement. We use laser generation and detection, and compare results with finite element method (FEM) models. We also investigate frequency enhancements for angled and branched defects, to highlight further potential measurement techniques when using scanning laser detection.

Wedge-shaped samples can be used as a model of acoustic interactions with samples ranging from ocean wedges, to angled defects such as rolling contact fatigue, to thickness measurements of samples with non-parallel faces. We present work on laser generated ultrasonic waves on metal samples; one can measure the dominant Rayleigh-wave mode, but longitudinal and shear waves are also generated. We present calculations, models, and measurements giving the dependence of the arrival times and amplitudes of these modes on the wedge apex angle and the separation of generation and detection points, and hence give a measure of the wedge characteristics.

A full matrix capture technique is presented that allows for real-time imaging through a non-planar surface where the geometry is known. For an identified geometry the point of incidence at the refractive interface is calculated using Fermat’s principle and iterative techniques for each possible transducer position which is pre-processed ahead of the inspection. This information is combined with the transducer’s encoded position during the inspection process with post-processing of ultrasonic data performed over the graphic processing unit. This is shown to allow for rapid imaging of ultrasonic data by firstly reducing the need to auto-focus through the media, and secondly by exploiting the parallelisation power of the graphic card. To demonstrate this, a linear array transducer was mounted to a Perspex wedge where the technique was applied to generate ultrasonic imagery of side drilled holes through a curved surface. This is shown to offer significant performance over traditional full matrix imaging with low implementation and development costs.

Standard test samples typically contain simulated defects such as slots machined normal to the surface. However, real defects will not always propagate in this manner; for example, rolling contact fatigue on rails propagates at around 25 degrees to the surface, and corrosion cracking can grow in a branched manner. Therefore, there is a need to understand how ultrasonic surface waves interact with different crack geometries. We present measurements of machined slots inclined at an angle to the surface normal, or with simple branched geometries, using laser ultrasound. Recently, Rayleigh wave enhancements observed when using the scanning laser source technique, where a generation laser is scanned along a sample, have been highlighted for their potential in detecting surface cracks. We show that the enhancement measured with laser detector scanning can give a more significant enhancement when different crack geometries are considered. We discuss the behaviour of an incident Rayleigh wave in the region of an angled defect, and consider mode-conversions which lead to a very large enhancement when the detector is close to the opening of a shallow defect. This process could be used in characterising defects, as well as being an excellent fingerprint of their presence.

The use of lasers for generating and detecting ultrasound is becoming more established in non-destructive testing. However, there is still scope in developing the techniques to fully realise the benefits of non-contact measurements. One application is the detection of surface defects in metals; for example, rolling contact fatigue in rails, and surface cracking on billets or plates. We present measurements using a pulsed Nd:YAG laser to generate surface ultrasonic waves and an interferometer to detect the surface displacement on the sample, and investigate the interaction of Rayleigh or Lamb waves with surface defects. Signal enhancement in the near-field is observed for Rayleigh waves when either the generator or detector is close to a defect. For a scanned detector measurement, enhancement is observed due to constructive interference of the incident and reflected waves. For a scanned generator measurement, the change in generation conditions when the laser is over the defect also lead to an enhancement. In measurements of plate samples we observe similar enhancement effects whereby higher order modes are observed when the laser is above a defect. We discuss the implications of signal enhancements for detecting and characterising surface cracking.

Full matrix capture allows for the complete ultrasonic time domain signals for each transmit and receive element of a linear array probe to be retrieved. While it is more common to use full matrix capture to post-process data to allow for electronic steering, focusing and imaging after the initial inspection processes, due to the data-acquisition and performance limitations of the focusing algorithms real-time inspection systems are not yet common place.This paper investigates several algorithm optimisation techniques utilising standard in-expensive PC architecture with parallelisation undertaken by the graphic processing unit. This approach is further combined with several other software engineering optimisation techniques including threaded data-capture, the use of look-up tables and half-matrix implementation to produce a real-world inspection scenario for benchmark and performance analysis.Experimental results are presented indicating that high frame rates inclusive of data-acquisition and image render are achievable with 32 active transmit and receive elements. This approach is shown to offer significant performance advantages with low implementation and development costs. Crown Copyright (C) 2012 Published by Elsevier Ltd. All rights reserved.

Enhancement of the Rayleigh wave signal amplitude at a surface defect, due to interference of incident, reflected and mode converted waves, has been reported by several authors, and it has been suggested that this could be used as a fingerprint of the presence of such cracking. The scanning laser line source technique in particular, where signal amplitude is enhanced as the laser generating the Rayleigh waves is in the region of a surface defect, has been reported as a suitable detection tool. However, the previous work has looked primarily at defects propagating normal to the surface, which may not always be a suitable approximation, and the enhancement measured when a detection laser rather than a generation laser is near a crack may, in some cases, be more significant. This work explores near field effects for both laser generation and laser detection points near a defect, and compares the enhancements for defects which are angled relative to the surface. We use a combination of finite element method models and experimental results, and probe enhancements of both the amplitude and frequency signals, and show that scanning the detection point may be a better method for locating surface defects if they are inclined at an angle to the surface.

Non-linear enhancements of ultrasonic surface wave amplitude and frequency have been observed when an incident wave interacts with a surface defect. Previous measurements of surface wave interactions with defects have considered only those that are inclined normal to the surface. Here, the enhancement effects have been studied in aluminium samples with machined slots of fixed length and of varying angle to the horizontal; the degree of enhancement was studied as a function of defect angle using both a scanning laser source, and a scanning laser detector. An automated scanning system has been developed for use with the detector, an IOS two-wave mixer interferometer, capable of measuring the out-of-plane surface displacement on rough surfaces. B-scans, consisting of many A-scans stacked together, were used to identify wave modes present in the near field, the arrival times of which are dependent on the angle of the defect. The observed enhancement is caused by superposition of the incident Rayleigh wave with reflected and mode converted waves, thereby making it angle dependent.

Non-destructive testing is an important technique, and improvements are constantly needed. Surface defects in metals are not necessarily confined to orientations normal to the sample surface; however, much of the previous work investigating the interaction of ultrasonic surface waves with surface-breaking defects has assumed cracks inclined at 90° to the surface. This paper explores the interaction of Rayleigh waves with cracks which have a wide range of angles and depths relative to the surface, using a non-contact laser generation and detection system. Additional insight is acquired using a 3D model generated using finite element method software. A clear variation of the reflection and transmission coefficients with both crack angle and length is found, in both the out-of-plane and in-plane components. The 3D model is further used to understand the contributions of different wavemodes to B-Scans produced when scanning a sample, to enable understanding of the reflection and transmission behaviour, and help identify angled defects. Knowledge of these effects is essential to correctly gauge the severity of surface cracking.

The behaviour of sound waves interacting with wedges has attracted interest from researchers in geophysics and non-destructive testing. We consider here the near-field behaviour of Rayleigh waves incident on wedges and surface-breaking defects which propagate at an angle to the surface, such as rolling contact fatigue on rails. It has been shown that, for a detection point on the edge of the crack tip, a very large signal enhancement is observed for shallow angles. We explain this behaviour through considering the effect of the defect geometry, with changes in the frequency·thickness product leading to mode-conversion of the incident Rayleigh wave.

This paper explores some effects that occur when using laser ultrasound to scan defective samples. Surface defects can often propagate at an angle to the surface; however, for calibration, slots machined normal to the surface of the sample are typically used. Several interesting angle-dependent effects are observed when Rayleigh waves interact with angled surface defects, and are explored here using measurements and models for a scanning laser detector (SLD) or scanning laser line source (SLLS) scanned across the defect. Reflection and transmission coefficients are calculated for different crack angles and lengths. Additionally, interesting angle-dependent effects are observed in the Rayleigh wave amplitude and frequency enhancements in the near field when using SLD or SLLS.

A pulsed Nd:YAG laser with an approximately Gaussian beam shape is directed onto the surface of an aluminium sheet at an energy density below which damage by laser ablation occurs, generating Lamb waves in the sheet. The laser beam is raster scanned across the surface of the sample. The Lamb waves travel radially outwards from the generation point and are detected some distance away by an electromagnetic acoustic transducer with sensitivity to in-plane displacements of the sheet. A number of static EMATs are located around the edges of the sheet, some distance from the generation point. The presence of a crack-like defect on the sheet can be detected by either a sudden change in the ultrasonic waveform or by an enhancement in the frequency content of the waveform when the laser beam illuminates directly onto the crack.

Surface ultrasonic waves have been shown to have many uses in non-destructive testing, in particular for gauging the depth of surface defects. Much of the previous work has assumed that these defects are oriented normal to the surface. However, this is not always the case; for example, rolling contact fatigue in rails propagates at an angle of around 25 to the surface, and this angle may affect the characterisation. We present results using non-contact ultrasonic methods to generate and detect ultrasound on samples with a range of defect angles, and compare these with finite element method (FEM) models. We use both electromagnetic acoustic transducers (EMATs) and laser ultrasound. The depth calibration when measuring ultrasound transmission is considered, and what affect the angle of a defect has. Several other methods of characterising crack depth and angle are also discussed, including the arrival times of reflected and mode-converted waves, the delay in the transmission of the high-frequency Rayleigh wave, and the enhancement of the signal at the defect in both the in-plane and out-of-plane components.

Electromagnetic acoustic transducers (EMATs) have been used to generate and detect Rayleigh waves in order to identify surface cracking in aluminium bars and rails. B-scans produced during scans of samples were used to determine the presence of surface defects. Additionally, the differences between signal enhancements due to wave interference at the crack produced by normal (900) and angled cracks in the B-scans were used to classify samples in order to decide an appropriate depth calibration curve for depth estimation. Classification was done using an image processing algorithm that selected the best features for classification, and used these to identify similar patterns in unclassified B-scans.

In this paper we exploit the interaction of ultrasonic surface waves with surface cracks in order to extract defect characteristics. An experimentally validated computer model was developed to simulate laser generated surface waves interacting with defects of several depths and angles. The Rayleigh reflection coefficient vs. crack angle and depth was explored. Amplitude and frequency behaviour at the defect interface in aluminium samples was also considered. Additionally, the phase component of the FFT was used to extract similar information. Finally, we explore the time‐frequency behaviour using the Wigner transform.

One of the contributing factors to graphite degradation is oxidation. In order to detect this type of degradation, samples with different levels of oxidation have been studied. Longitudinal velocity measurements were carried out on these samples, using laser ultrasound techniques. These techniques have the advantage that they are non-contact, and therefore do not perturb the sample under investigation. From these measurements, the Young’s modulus was estimated, based on a typical Poisson’s ratio of 0.20. This was performed for several cylindrical samples; some samples had parallel microstructure orientation in the through-thickness measurement direction, while others had perpendicular orientation. Both orientations had very different longitudinal velocities, with corresponding variations in Young’s modulus. Degradation was assessed for both cases of parallel and perpendicular oriented samples. Furthermore, in these perpendicular oriented samples, velocities were dependent on sample rotation about the cylindrical axis. Additional measurements were taken at 90° orientations to confirm this feature. The data positively show polarization of longitudinal velocity components. This may be the first time that acoustic birefringence has been observed in high porosity graphite.

This paper describes two applications of laser/ electromagnetic acoustic transducer (EMAT) systems. The first application uses laser-generated ultrasound for the characterisation of rear surface artificial defects (vertical slots) in metal samples. An EMAT sensitive to in-plane motion was used to detect these ultrasonic waves. B-scan images were used to visualise any changes of interrogating waves due to defects. These images were generated as the sensor head was moved along the material's surface, forming a 2D intensity profile that revealed changes in the presence of a defect. The defects produced distinctive parabolic features in the images from the time-of-flight diffraction by shear waves. The paper presents a new method based on a generalized Radon transform (GRT) to automatically determine the depth and location of back-surface artificial defects from B-scan images. Experimental evidence validated predicted results for the case of 1.5-3.5 mm-deep defects. The second application uses Rayleigh waves to quantify inhomogeneities in metal alloys using a transient Rayleigh pulse detected with an eight-element EMAT-array. The array spacing established sensor separation, so that velocity measurements were performed independent of the source-to-detector separation. Eliminating this distance, which would normally lead to systematic errors, has produced a measurement system capable of measuring velocities in metal with a precision of 99.9 %. As an example, variations in the Rayleigh wave velocity have characterised the inhomogeneities in rolled bars of aluminium.