Jorge F. Cruza

Automatic dynamic depth focusing for NDT

Auto-focusing along with dynamic depth focusing (DDF) would be very valuable to inspect arbitrarily shaped parts when operating with wedges or with other coupling media to avoid the burden of computing and setting the correct focal laws while still getting the best possible resolution at all depths. This work proposes a three-step procedure to perform the auto-focusing function with DDF in real time. First, the part geometry is estimated by the first echo time-of-arrival following one of several possible strategies: pulse-echo, pitch-catch, or plane wave. These are analyzed with regard to their performances and acquisition time, giving closed formulae to get the coordinates of interface points. After a curve fitting and extrapolation process, a virtual array that operates in a homogeneous medium is computed, avoiding the complications of refraction at the interface and allowing operation with already known focusing hardware. This hardware is initialized with the set of focusing parameters adapted to the estimated probe- part geometry, and ensures that all received samples are in focus. Using a standard computer, the auto-focusing procedure currently takes about 2 s to perform. Experiments carried out under different conditions validate the proposed technique.

New method for real-time dynamic focusing through interfaces

In nondestructive evaluation (NDE) a coupling medium (wedge) is frequently inserted between the array probe and the object being evaluated. In this situation, focal law computing is complicated by the refraction effects at the interface. Furthermore, there are not known techniques to perform dynamic focusing by hardware in these conditions. This work addresses these problems by following a two-step procedure. First, a virtual array that operates in a single medium with nearly equivalent time-of-flight to the foci is obtained. Then, simple hardware is proposed to perform dynamic focusing in real-time. It operates with arrays of any geometry as required by the virtual array in presence of arbitrarily shaped interfaces. The paper describes the theory and evaluates the timing errors of the approximations made. These errors are low enough to allow use of the new technique in most NDE and some specific medical applications. The new technique is validated by simulation and experimentally.

Auto-focused virtual source imaging with arbitrarily shaped interfaces

This work presents a new method, named autofocused virtual source imaging (AVSI), for synthetic aperture focusing through arbitrarily shaped interfaces with arrays. First, the shape of the component surface is obtained by timeof- flight (TOF) measurements. Then, a set of virtual source/ receivers is created by focusing several array subapertures at the interface normal incidence points. Finally, the synthetic aperture focusing technique (SAFT) is applied to the received signals to generate a high-resolution image. The AVSI method provides several advantages for ultrasonic imaging in a twomedia scenario. First, knowledge of the probe-part geometry is not required, because all information needed for image formation is obtained from a set of ultrasonic measurements. Second, refraction complications in TOF calculations are avoided, because foci at the interface can be considered as virtual source/ receivers, and SAFT can be performed in the second medium only. Third, the signal-to-noise ratio is higher than with synthetic aperture techniques that use a single element as emitter, and fourth, resolution is higher than that obtained by phasedarray imaging with the same number of active elements, which reduces hardware complexity. The theoretical bases of the method are given, and its performance is evaluated by simulation. Finally, experimental results showing good agreement with theory are presented.

Air-coupled ultrasound inspection of complex aluminium-CFRP components

Composite Overwrapped Pressure Vessels (COPVs) are widely used to contain high-pressure fluids. COPVs consist of a thin metallic liner covered with a composite material based on structural fibres. The main function of the liner is to avoid direct contact between composite and chemical agents which can degrade the resin matrix. In the filament winding process for COPV production, a metallic core is wrapped with impregnated carbon filaments. After that, the matrix can be cured at ambient temperature without use of an autoclave. The COPVs structural integrity can be assured using suitable designs, controlled manufacturing processes and effective Non Destructive Evaluation (NDE) methods. However, the ultrasound inspection of these components is challenging by several factors. Because no vacuum compaction process is applied, the sound attenuation in the laminate is high. Moreover, the strong reflection at the liner-fibre interface avoids inspecting from the metallic-side using conventional techniques. This work explores the possibility of inspecting COPVs by air-coupled ultrasound after the manufacture process.

High resolution Autofocused Virtual Source Imaging (AVSI)

Synthetic aperture focusing technique with arrays (SAFT) provides improved lateral resolution but, when a coupling medium is between the probe and the part, refraction complicates time-of-flight computations. This slows down the imaging process since there are not general formulae for arbitrarily shaped interface. The approach of the AVSI proposed method is to generate, by conventional delay and sum beamforming, a set of foci at different points of the interface. These act as virtual sources and SAFT processing can be carried out in the second medium only, avoiding refraction complications. Furthermore, an automated process for locating the virtual sources from ultrasound measurements and to compute the adapted focal laws in emission and reception is proposed, so knowledge of the part shape geometry is not required. Theoretical bases of the Autofocused Virtual Source Imaging method (AVSI) are presented, along with experimental data that confirms the expected performance.

Real time autofocusing hardware for ultrasonic imaging with interfaces

A hardware architecture for fast imaging (multi-beamforming) with auto-focusing capabilities is presented. Auto-focusing is carried out without the knowledge of probe-part geometry, which is obtained by ultrasound measurements. The proposed technique and hardware achieves a high throughput that allows autofocus in real-time (for instance, in part scanning applications). The paper yields some figures of performance and other preliminary results.

Automatic dynamic focusing through interfaces

An interface between the coupling medium and the inspected part is frequently found in Non Destructive Testing (NDT). When phased array technology is used, the focal laws must be obtained considering the refraction at the interface, which requires computationally intensive iterative procedures since no closed formulae exist for the general case. This work presents a new two-step approach. In the first step, the two propagation media scenario is converted into a single homogeneous medium by computing a virtual array with nearly equivalent flight times to the foci. In the second step, a focusing hardware, conveniently initialized, evaluates in real-time the sampling instants that correspond to the focal laws. Focusing errors are small enough to validate the new technique for a wide range of applications. In fact, for active apertures currently used in NDT, the resolution and dynamic range are almost not affected as it is experimentally shown. Furthermore, focal law computing time is dramatically reduced by evaluating a few parameters instead of focusing delays for every output sample, element and steering direction. The instrument focal law storage requirements become significantly reduced as well.

Total Focusing Phased Array

Phased Array (PA) uses a single focus in emission and Dynamic Depth Focusing (DDF) in reception, while Total Focusing Method (TFM) focuses every image pixel in emission and in reception. By contrast, TFM lacks the A-scan information, which is standard in PA and can be very useful for evaluation. Another issue common to both techniques is the difficulty associated with two media propagation and unknown interface geometry. A further TFM drawback is the limited image extent due to memory and time consumption. This work addresses these problems by proposing the new Total Focus Phased Array (TFPA) technique that combines the advantages of PA and TFM. TFPA builds A-scan lines as in PA, but focused in emission and in reception at all depths. This process is achieved in real-time by N+1 focusing hardware blocks that do not limit the length of every A-scan. Furthermore, TFPA can operate with the Virtual Array concept that avoids the problems of refraction in two propagation media applications. TFPA has many advantages: A-scan information is preserved, image is focused in emission and in reception, moderate complexity with real-time capability, no refraction issues and no limit on the depth or the number of pixels/samples in the image.

Spatial Compounding of ARFI images in an automated breast imaging system

Automated ultrasound breast imaging could overcome most of the limitations of conventional echography, ensuring repeatability and better patient follow-up. Several research groups have included multi-modality in these kind of systems, obtaining velocity and attenuation maps. This work addresses the implementation of the Acoustic Radiation Force Imaging (ARFI) technique in a Full Angle Spatial Compounding (FASC) automated breast imaging system to obtain tissue elasticity images. Results demonstrated that ARFI imaging can be applied efficiently without direct contact between transducer and tissue (water coupling), and that full angle composition improves image quality. Furthermore, detection capability of small structures improves with regard to images obtained from a single array location, in which image texture can mask their presence.

Total Focusing Method With Virtual Sources in the Presence of Unknown Geometry Interfaces

Auto-focused virtual source imaging (AVSI) has been recently presented as an alternative method for synthetic aperture focusing through arbitrarily shaped interfaces with arrays. This paper extends the AVSI concept to the case of the total focusing method (TFM-AVSI) using several virtual receivers for each virtual source. This approach overcomes the known contrast limitation of AVSI, while preserving the advantage of performing synthetic focusing in the second medium only [no time-of-flight (TOF) calculations through the interface]. In contrast, equipment with more active channels must be used to digitalize the signals received by all the array elements after each focused emission. When compared with the conventional TFM, the proposed method reduces the processing complexity of the most time consuming task: TOF calculation in the presence of interfaces. This improvement could lead to more efficient real-time implementations of the TFM in non-destructive testing applications where water immersion or flexible wedges are used. In this paper, the mathematical formulation for the new method is given, accounting for the surface slope and the array angular sensitivity. Its performance is evaluated by numerical simulation, experimentally and compared with AVSI and the conventional TFM. It was found that the TFM-AVSI achieves the same resolution and contrast as that of the TFM, although it shows a wider blind zone below the interface due to focusing with normal incidence.