Examples of Specific ApplicationsSound & Bright is commited to the development of high-precision, practical systems that overcome both research and industrial challenges.
Detection of Through-Transmission Signal Generated by Thermoelastic Laser Pulses
To demonstrate bi-componenet detection capability using the Tempo 2D System to scan an aluminum sample.
An example of in-plane and out-of-plane detection is shown below. Figure 1 describes a through transmission experiment. Generation was achieved in a thermoelastic regime with a Nd:YAG pulsed laser. The laser beam was focused along a line with a cylindrical lens. The detection was carried out on the opposite side of the sample with a bandwidth between 20 kHz and 20 MHz. The sample was a 12.7 mm thick aluminum plate. The laser ultrasonic receiver scanned along a 50 mm line. In a first experiment, we used a sample free of defects (no blind holes), while in the second, two seperate blind holes (as shown Figure 1) were introduced.
The B-scan results for both in-plane and out-of-plane displacements are shown in Figure 2-A. The direct P-wave (P) and S-wave (S) arrivals at their respective arrival time 2 μs and 4 μs are clearly visible, along with multiple reflected and reflected/converted waves. As expected, the thermoelastic source generates a stronger S than P-wave. For these measurements, the calibration coefficient is 100mV/nm for both in-plane and out-of-plane outputs. After 12μs, some reflections from the sample edges are also visible on both in-plane and out-of-plane measurements. Moreover, as expected, the in-plane B-scan signal is asymmetrical around the epicenter, whereas the out-of-plane B-scan signal is symmetrical around the epicenter.
Two defects (blind holes as described in Figure 1) were introduced in the sample and the measurements were repeated. The result comparing the B-scans, without and with defects, is shown in Figure 2. Reflection / diffraction from the two defects are visible on both signals, but they are clearly identifiable on the in-plane B-scan.
Lamb Mode Resonance Measurement on Thin Plate
To demonstrate low frequency Lamb wave detection capability using the Tempo 2D System to scan an aluminum sample.
In this next experiment, the 12.7 mm thick aluminum plate is replaced with a 1 mm thick aluminum sheet. The generation was carried out with a line-focused pulsed laser beam. The measured in-plane direction is along the propagation direction. Figure 3-A illustrates low-frequency Lamb waves detected with the Tempo 2D after propagation along 50 mm. The detection bandwidth is the same as the previous B-scans. The signals were low-pass filtered with a cut-off frequency set at 1MHz, in order to only visualize the lower frequencies for the first set of measurements.
Figure 3-B describes the result of this experiment. The Lamb wave mode A0 is clearly visible after 16μs on both graphs. The S0 mode is clearly present on the in-plane signal, but hardly distinguishable on the out-of-plane. The graph enhances the fact that the symmetrical S0 mode has a stronger in-plane component which propagates faster than the asymmetrical A0 mode. Moreover, we can clearly see that the asymmetrical A0 carries more energy than the S0 mode. Finally, the A0 mode presents a phase shift of 90° between the in-plane and out-of-plane components, which justifies its asymmetrical property.
A second set of measurements is shown on Figure 4, with detection at the epicenter. Here the data are high-pass filtered, showing only the frequency above 1MHz.
Strong resonances are detected. Some Lamb modes exhibit an anomalous behavior at frequencies where the group velocity vanishes while the phase velocity remains finite. The zero-group velocity (ZGV) leads to sharp continuous resonance and ringing effect. Figure 4 shows the fast Fourier transform computed on the first 50μs of the signals. The spectrum of the out-of-plane signal (Figure 4-A) and the in-plane signal (Figure 4-B) clearly show the resonance of the S1 mode and of the A2 mode as described by Clorennec et al in their paper Laser Impulse Generation and Interferometer Detection of Zero Group Velocity Lamb Mode Resonance (Appl. Phys. Lett. 89, 2006). The resonance of the A2 Lamb mode corresponds to the thickness shear resonance (F_2.d=3.V_s/2), where VS is the shear wave velocity and d is the thickness.
Non Destructive Evaluation on moving samples
To demonstrate high-sensitivity measurement capability on a fast moving sample using a Quartet System.
Thanks to a short response time, the Quartet features the ability to perform highly sensitive measurements on moving samples. To demonstrate this capability, we performed experiments on a rotating disc composed of aluminum, 9 mm thick with a diameter of 130 mm. One side had a natural (reflective) finish while the other was sandblasted to create a uniform light-scattering surface.
The signal is first generated by a piezoelectric mirror continuously exiting at f=1.6MHz in the reference path. The signal amplitude is measured at 2f=3.2 MHz (Fig. 1A), because the rectification generates the second harmonic. There is almost no influence on the reflective side below 2 m/s, and then the signal drops slightly (about 5 dB) between 2 and 3 m/s. On the scattering surface, the signal begins 5dB lower than the reflective surface because less light is collected. The drop appears at a lower linear speed, but the signal is still detected. An example of pulse-echo LBU signals recorded with an offset between generation and detection is shown Fig. 4B.
Figure 2 shows B-scans performed in a pulse-echo experiment on the scattering side with either the sample at rest or moving and scanning of the generation laser source. The linear speed at the point of detection was 2 m/s. The first three reflected P-waves were detected, as well as the surface wave, the first reflected shear wave and the converted wave (SP). This demonstrates the ability of the interferometer to process a signal even on rough,fast moving samples.
Non Destructive Evaluation on High-Temperature Ceramics
To demonstrate measurment capability on a highly scattering surface at high-temperature using the Quartet System.
Here we demonstrate the ability to carry out measurements on a highly scattering surface, at high temperature. The sample is a high-temperature ceramic brick that is used for building walls containing molten glass, in industrial glass ovens. During manufacturing, one side of the ceramic reaches a temperature of 1500oC and can be in contact with molten glass. The piece of ceramic measured was part of a door of a high-temperature laboratory oven in which the inside temperature could reach 1500oC. The sample dimensions were: 5cm (thickness) x 7cm x 10cm. The setup was a pulse-echo experiment with an offset between generation and detection. The generation was line focused to about 2 mm x 10 mm. The B-scan was recorded by scanning the generation. In this experiment, we used a shorter stand-off distance of 20cm in order to maximize the amount of light back-scattered by the rough and highly diffusive surface. Each piece of data was averaged over 100 acquisitions.
Strong ultrasonic scattering and multiple reflections on the edge of the small sample make the identification of the waves difficult. An example of LBU signal is shown Figure 3. The time scale in Figure 3 has been shifted, with t=0 corresponding to the arrival of the P-wave reflected through the thickness. The B-scan recorded by scanning the position of the generation line facilitates the identifications of the various waves. Figure 4 shows the B-scans recorded at three different temperatures: 800, 1300 and 1500oC. The surface wave (R) as well as the reflected surface wave (RR), the direct compression wave (P-direct), the reflected compression wave (PP) and the reflected-converted wave (PS &SP) are clearly identifiable.
Using laser interferometry to investigate strong multiple scattering of ultrasound near the Anderson transition in a three-dimensional “mesoglass”.
Research By: Professor John Page, Distinguished Professor of the University of Manitoba
Measure ultrasonic displacements on the surface of a brazed aluminum beads sample directly using a laser interferometer to explore ultrasonic wave propagation.
These samples scatter ultrasonic waves very strongly and are of interest because they may exhibit the unusual phenomenon of Anderson localization.
In strong scattering media, the propagation of waves can often be described using the diffusion approximation. When the disorder increases, Anderson localization may occur. Scaling theory tells us that there is a transition from diffusive to localized behavior in three dimensional systems. This different behavior has been demonstrated by John Page’s group in similar samples. In the diffusive regime, the wave energy spreads out from the source point, but in the localized regime, the wave spreading is cut off and one can define the localization length from the width of the spatial profile.
In the simplest case, the multiply scattered waves undergo random walks through the sample and energy transport is diffusive; then, the width of the spatial profile grows as the square root of time, and can be used to measure the wave diffusion coefficient. The best way of directly measuring this effect is to monitor the time dependence of the transverse width of the spatial profile on the sample surface opposite the source point.
The aluminum bead mesoglass samples are especially interesting physically because the disorder is so strong that instead of spreading diffusively without limit, the time dependent growth of the transverse profile is cut off by interference effects, and the waves remain trapped or localized inside a finite region of the sample centred near the source.
The sample is placed in air so there are virtually perfectly reflecting boundaries with a large acoustic mismatch. Luckily, this simplifies the theoretical interpretation. This plot shows typical data acquired when a short pulse of ultrasound is incident on the opposite side of the sample. Displacement near this red circle. The measured voltage from the laser interferometer can be converted to the actual displacement directly in a very convenient way. The normal displacements were measured over a 2 by 2 “ area with spatial resolution of 0.008”. The surface structure of the modal displacements is well-captured by the performed scan.
Here, the evolving field profile is visible. The spatial resolution of the scan is good enough to display the smooth variation of the field and the emerging field profile as a function of time is visible on the surface of the sample. The magnitude of FFT of the normal displacement at different frequencies is also shown.
Multicomponent laser interferometry for reduced scale physical (seismic) modeling
Research By: Raphael Valensi, Donatienne Leparoux, Olivier Durand
To reproduce seismological field experiments on a laboratory scale through modeling.
To use multi-component inspection to reconstitute the complete ultrasonic field.
The principle objective of reduced-scale-modeling is to reproduce field experiments on a laboratory scale. When compared to real (or field scale), higher frequencies are used in order to maintain a constant number of propagated wavelengths over smaller (lab scale) distances. Depending on the geophysical context, the scaling factors between real and reduced scale modeling can vary between 103 and 108. Advantages of reduced scale modeling in Seismology are:
– Seismological Modeling is difficult to compute.
– Experiments are conducted in a controlled environment, and therefore reproducible
– It represents an alternative way of providing data to test imaging methods
More generally, there is great interest within the geophysical community for multicomponent measurements. An ultrasonic measurement bench was designed to reproduce an active seismic experiment:
– The seismic source is a piezoelectric transducer (moving according to 1 DoF).
– The propagation media is a model (polymers, metal … ) lying on an optical table.
– The sensor is a Tempo 2D laser interferometer measuring the particle displacement (moving according to 2 DoF).
Frequency range : [20 kHz; 900 kHz].
A reduction of the rigging effect was achieved using a “recompression” process in the frequency domain from an isolated wavelet: this process does not introduce bias in the spectral ratios and allows for the separation of waves by time windowing.
Observation of the polarization of the measured Rayleigh wave in the time domain: