No abstract

This work proposes a novel hybrid signal processing technique to extract information on disbond-type defects from a single B-scan in the process of non-destructive testing (NDT) of glass fiber reinforced plastic (GFRP) material using ultrasonic guided waves (GW). The selected GFRP sample has been a segment of wind turbine blade, which possessed an aerodynamic shape. Two disbond type defects having diameters of 15 mm and 25 mm were artificially constructed on its trailing edge. The experiment has been performed using the low-frequency ultrasonic system developed at the Ultrasound Institute of Kaunas University of Technology and only one side of the sample was accessed. A special configuration of the transmitting and receiving transducers fixed on a movable panel with a separation distance of 50 mm was proposed for recording the ultrasonic guided wave signals at each one-millimeter step along the scanning distance up to 500 mm. Finally, the hybrid signal processing technique comprising the valuable features of the three most promising signal processing techniques: cross-correlation, wavelet transform, and Hilbert–Huang transform has been applied to the received signals for the extraction of defects information from a single B-scan image. The wavelet transform and cross-correlation techniques have been combined in order to extract the approximated size and location of the defects and measurements of time delays. Thereafter, Hilbert–Huang transform has been applied to the wavelet transformed signal to compare the variation of instantaneous frequencies and instantaneous amplitudes of the defect-free and defective signals.

In this paper, a novel 2D analytical model based on the Huygens’s principle of wave propagation is proposed in order to predict the directivity patterns of contact type ultrasonic transducers in the generation of guided waves (GWs). The developed model is able to estimate the directivity patterns at any distance, at any excitation frequency and for any configuration and shape of the transducers with prior information of phase dispersive characteristics of the guided wave modes and the behavior of transducer. This, in turn, facilitates to choose the appropriate transducer or arrays of transducers, suitable guided wave modes and excitation frequency for the nondestructive testing (NDT) and structural health monitoring (SHM) applications. The model is demonstrated for P1-type macro-fiber composite (MFC) transducer glued on a 2 mm thick aluminum (Al) alloy plate. The directivity patterns of MFC transducer in the generation of fundamental guided Lamb modes (the S0 and A0) and shear horizontal mode (the SH0) are successfully obtained at 80 kHz, 5-period excitation signal. The results are verified using 3D finite element (FE) modelling and experimental investigation. The results obtained using the proposed model shows the good agreement with those obtained using numerical simulations and experimental analysis. The calculation time using the analytical model was significantly shorter as compared to the time spent in experimental analysis and FE numerical modelling.

No abstract

No abstract

Multi-wire ropes are widely used as load-carrying constructional elements in bridges, cranes, elevators, etc. Structural integrity of such ropes can be inspected by using non-destructive ultrasonic techniques. The objective of this work was to investigate propagation of ultrasonic guided waves (UGW) along composite multi-wire ropes in the cases of various types of acoustic contacts between neighboring wires and the plastic core. The modes of UGW propagating along the multi-wire ropes were identified using modelling, the dispersion curves were calculated using analytical and semi-analytical finite element (SAFE) techniques. In order to investigate the effects of UGW propagation, the two types of the acoustic contact between neighboring wires were simulated using the 3D finite element method (FE) as well. The key question of investigation was estimation of the actual boundary conditions between neighboring wires (solid or slip) and the real depth of penetration of UGW into the overall cross-section of the rope. Therefore, in order to verify the results of FE modelling, the guided wave penetration into strands of multi-wire rope was investigated experimentally. The performed modelling and experimental investigation enabled us to select optimal parameters of UGW to be used for non-destructive testing.

Hidden corrosion defects can lead to dangerous accidents such as leakage of toxic materials causing extreme environmental and economic consequences. Ultrasonic guided waves showed good potential detecting distributed corrosion in pipeline networks at sufficiently large distances. To simplify signal analysis, traditional guided wave methods use low frequencies where only fundamental modes exist; hence, the small, localized defects are usually barely detectable. Novel techniques frequently use higher order guided wave modes that propagate around the circumference of the pipe and are more sensitive to the localized changes in the wall thickness. However current high order mode guided wave technology commonly uses either non-dispersive shear modes or higher order mode cluster (HOMC) waves that are mostly sensitive to surface defects. As the number of application cases of high order modes to corrosion detection is still limited, a huge potential is available to seek for other modes that can offer improved resolution and sensitivity to localized corrosion type defects. The objective of this work was to investigate higher order modes for corrosion detection and to determine the most promising ones in sense of excitability, leakage losses, propagation distance, and potential simplicity in the analysis. The selection of the proper mode is discussed with the support of phase and group velocity dispersion curves, out of plane and in plane distributions over the thickness and on surface of the sample, and leakage losses due to water load. The analysis led to selection of symmetric S3 mode, while the excitation of it was demonstrated through finite element simulations, taking into account the size of phased array aperture and apodization law and considering two-sided mode generation. Finally, theoretical estimations were confirmed with experiments, demonstrating the ability to generate and receive selected mode. It was shown that S3 mode is a good candidate for corrosion screening around the circumference of the pipe, as it has sufficient propagation distance, can be generated with conventional ultrasonic (UT) phased arrays, has sufficiently high group velocity to be distinguished from co-existing modes, and is sensitive to the loss of wall thickness.

Fiber-reinforced composite laminates are being increasingly used in various engineering components in the sectors of aerospace and green energy. Due to impacts throughout the service life of the structure, matrix breakage and delaminations significantly altering the structural integrity of the laminate can occur. Hence, robust guided wave structural health monitoring systems are required to ensure continuous safety of engineering structures. In this paper, the ultrasonic method based on the analysis of A0 mode reflecting within the defected area has been proposed to extract the length and the depth of the delamination-type defect. The technique proposed in this study extracts the depth of the damage by analyzing the magnitude variations of direct A0 mode which are caused by the difference of wave velocities in the upper and lower sub-laminates. This results in an altering and frequency-dependent forward-scattered amplitude of direct A0 mode. Furthermore, the proposed approach uses previously obtained information about the depth of the defect, which allows for the determination of the phase velocities of A0 and S0 modes in the upper and lower sub-laminates. As a result, the accuracy of the damage length estimation is increased. The performance of the proposed method was proven with 2D and 3D numerical simulations and experiments on samples with artificial defects. The method validation results showed that the proposed method with some limitations is capable of extracting the length of delamination with an approximate error below 6%.