Ultrasonic detection of defects in strongly attenuating structures using the Hilbert–Huang transform

Abstract Non-destructive testing of polymer materials meets specific problems caused by a high attenuation of ultrasonic signals. When the expected flaws are close to the interfaces, the signals reflected by the flaws may be hidden in the tail of much stronger signals caused by front reflection from the sample and by interfaces between the adjacent layers. For improved detection of defects in plastic pipes, a new approach based on a combined application of non-linear deconvolution and the Hilbert–Huang transform is proposed. The first step of the algorithm is the elimination of strong ultrasonic echo signals reflected by interfaces. The second step is processing of ultrasonic signals using the Hilbert–Huang method. In order to improve visualization of defects we have proposed a new presentation of the Hilbert–Huang spectrum based on calculation of the product of the amplitude and the instantaneous frequency of analytic signal and displaying this product in a three-dimensional plot. The experimental investigations demonstrated a good performance of the proposed technique in the case of highly attenuating plastic pipes.