Evaluation of a multiple scattering filter to enhance defect detection in heterogeneous media

Abstract

Ultrasonic evaluation of coarse-grain materials generates multiple scattering at high frequency and large depth. Recent academic experiments with array probes showed the ability of a random matrix method [multiple scattering filter (MSF)] to reduce multiple scattering, hence improving detection. Here, MSF is applied to an industrial nickel-based alloy with coarse-grain structure. Two samples with average grain sizes 90 ± 60 μm and 750 ± 400 μm are inspected with wide-band 64-element arrays at central frequencies 2, 3, and 5 MHz. They contain cylindrical through-holes (1-mm radius) at various depths. The array transfer matrix is recorded and post-processed both in the flawless area and for eleven positions above each defect, which allows for a statistical analysis. MSF is compared with two conventional imaging techniques: the total focusing method (TFM) and the decomposition of the time-reversal operator (DORT). Several parameters to assess the performance of detection techniques are proposed and discussed. The results show the benefit of MSF, especially at high frequencies and for deep defects: at 5 MHz and 70 mm depth, i.e., more than three scattering mean-free paths, the detection rate for MSF ranges between 55% and 100% while it is found to be 0% both for TFM and DORT.

Quantification of the Effect of Multiple Scattering on Array Imaging Performance

“…The discrepancy between two estimations at the frequencies f > 3 MHz in the transition region z ≈ 15 mm is explained by the fact that the generalised image method is based on the averaged image properties. For the RRF sample the average single scattering intensity is always smaller (and, therefore, estimated multiple scattering rate is higher), compared to metric (18), where the maximum image amplitude is used. It needs to be stressed that the multiple scattering estimation method (10) is based entirely on the generalised image values outside of its main diagonal plane x T = x R , corresponding to the conventional 2D image.…”

“…The method is based on the specific phase property of the single-scattering contribution to discriminate singlefrom multiple scattered waves. This so-called multiple scattering filter (MSF) can be used to improve detection in random scattering media [17], [18], or to characterise the weakly scattering medium (human soft tissue) [6]. This technique was also applied as a metric to compare experimental measurements and FE simulations [19], and to investigate the effect of microstructural elongation in titanium alloys [20].…”

Matrix Approach of Seismic Imaging: Application to the Erebus Volcano, Antarctica

“…However, a systematic and quantitative analysis of the singular value spectrum is needed in order to assess whether the first two singular values are indeed associated with coherent reflectors at the focal depth or if they are false alarms due to speckle fluctuations. To that aim, a probabilistic approach based on random matrix theory is developed in the next section in order to discriminate potential artifacts (Aubry & Derode, , ; Shahjahan et al, ).…”

High-Density Polyethylene Pipe Butt-Fusion Joint Detection via Total Focusing Method and Spatiotemporal Singular Value Decomposition

“…After SVD processing, the defect signal is often distinguished from noise. This is due to the fact that the defect signal has a large energy share and corresponds to larger singular values, whereas noise corresponds to smaller singular values [23,24]. Zhang et al [25], in the detection of circular hole defects in noisy metallic materials, utilized high-order SVD to directly process FMC data.…”