Video ultrasonics by pulsed TV holography for nondestructive testing

Ultrasonic nondestructive testing (NDT) is a classical and powerful technology that has long been used for routine industrial inspections. By incorporating the latest scientific and technological advances, NDT methods have continuously been improved. With ultrasonic NDT, the object under inspection is excited with controlled sound waves that interact with flaws in the material and generate scattered waves, which are then detected at one or more locations (e.g., by measurement of the associated displacement or velocity). The detection stage can be implemented using many different contact (e.g., piezoelectric transducers) or noncontact methods (e.g., air-coupled ultrasonics), but optical detection of ultrasound is a relevant possibility that does not alter the measured quantity while providing high lateral resolution and other interesting features, such as the possibility to inspect small or inaccessible areas. In particular, full-field optical techniques (e.g., interferometry or speckle photography) can be used to make simultaneous measurements at all points in the field of view. This is therefore an attractive detection method because it permits inspection of large areas without the need for scanning.

Within the framework of full-field optical techniques, we have developed an ultrasonic NDT system with true transient analysis capability, which we call video ultrasonics by pulsed TV holography (VUPTVH).¹ Our technique combines ultrasonic probing and detection with TV holography²—also known as electronic speckle pattern interferometry (ESPI)—under pulsed laser illumination. With VUPTVH we are able to render high-quality movies that represent the spatiotemporal evolution of acoustic fields. Our research has focused on nonpolished plates and shell structures, which are commonly used in the fields of aeronautics and power generation.

Figure 1 illustrates our experimental setup, which is designed to record two displacement states of the object under test that are induced by the acoustic field. These measurements are subsequently processed to produce a digital map of the optical phase change between the states. This map is proportional to the out-of-plane component of the instantaneous acoustic displacement (on the order of 10nm). We use a series of digital maps, obtained under repeatable conditions and with an increasing delay between ultrasound generation and optical probing, to construct a video sequence that reproduces the spatiotemporal evolution of the acoustic field. For narrowband excitation, this data set has spatial periodicity (within each map) and temporal periodicity (at a pixel along the sequence), i.e., it carries both spatial and temporal information. We have therefore developed an additional evaluation stage that is based on a 3D Fourier transform of the video (see Figure 2).^3,4 The output of this stage is a new sequence of complex-valued maps, from which the acoustic amplitude, acoustic phase, and instantaneous acoustic displacement can be recovered independently with a substantial reduction in the noise component (see online video⁵).

Figure 1. Experimental setup and primary output (optical phase-change map, ΔΦ) of the video ultrasonics by pulsed TV holography. Nd:YAG: Neodymium-doped yttrium aluminum garnet.

Figure 2. The 3D Fourier transform evaluation method. In a typical experiment a sequence of 64 or 128 phase-change maps is used. t: Time.

We have shown experimentally that by using guided wave probing with VUPTVH it is possible to detect different types of defects in plate-like structures, including subsurface defects.^{1, 5} By processing the measured scattering sequences, we can analyze the potential to characterize the position, dimensions, and orientation of detected flaws. We therefore have to model scattering phenomena in the mid-high-frequency range (tens of ultrasonic wavelengths in the field of view) and for a large number of spatiotemporal samples (∼10⁸), which is a situation where many existing modeling approaches cannot be used.⁶ By using, as a first approximation, a simplified 2D scalar model to simulate the scattering of quasi-Rayleigh probing waves, it is possible to make quantitative characterizations of through-thickness reference defects (holes and slots, see online video⁷). We obtain a reasonable agreement (except for the backscattering amplitude) with our experimental results for both modulus and phase values, in near- and far-field zones simultaneously.⁸

The numerical backpropagation of acoustic wavefronts is another application that relies on the unique outputs of our system. We use longitudinal waves that propagate in thick metallic samples. The acoustic amplitude and phase of these waves are measured at the surface with VUPTVH, and we use a numerical implementation of the Rayleigh-Sommerfeld integral to reconstruct them at any depth within the sample (see Figure 3). We initially used this method to obtain the axial location of a subsurface defect embedded in a specimen.⁴ More recently we have used the technique to characterize the acoustic field of a transducer in a homogeneous solid medium and to assess the coupling between the sample and transducer (see online video⁹).¹⁰

Figure 3. Geometry for the measurement and backpropagation of longitudinal waves.

We have designed a new full-field optical technique for NDT applications. We plan to extend the capabilities of VUPTVH to include digital holography techniques that involve hybrid optical-numerical treatment of optical wavefronts. Further work is also required to verify the ability of our approach to assess a variety of guided wave types, defect topologies, and patterns. We are currently collaborating to develop a more rigorous approach, which is based on linear elasticity theory and state-of-the-art numerical methods, to make new comparisons with our experimental results.¹¹

Financial support for this work was gratefully received from the Spanish Research and Development National Plan, the European Regional Development Fund, Xunta de Galicia, and the University of Vigo.