Objective The positioning and even imaging of small defects inside metal workpieces can avoid unqualified workpieces. Au ultrasound imaging technology is based on the property of ultrasound which can penetrate opaque objects to obtain images of target objects, and is widely used in various fields. Traditional ultrasonic imaging data is mostly obtained by using ultrasonic transducers, such as piezoelectric transducers and electromagnetic acoustic transducers, to excite and detect ultrasonic waves on the surface of the material. However, these ultrasonic detection technologies have limitations in specific environments. On the other hand, laser ultrasonic nondestructive testing technology (NDT&E) can provide possibilities to overcome the shortcomings of traditional ultrasonic testing methods. Traditional laser ultrasonic defect detection methods include A-scan, B-scan, C-scan, etc. Most of these methods cannot accurately determine the size and location of the defect. Apart from this, the synthetic aperture focusing technology (SAFT) is widely used in the field of ultrasound imaging due to its ability of imaging of small defects and high imaging contrast. Our research combines the laser ultrasonic NDT&E with SAFT algorithm, which can realize imaging of small defect when it is used as the traditional ultrasonic transducer, and can also bring the advantages of laser ultrasound. The propagation and interaction with defects of laser ultrasound in metallic samples are numerically simulated by Comsol Multiphysics and the PSM-SAFT algorithm is compiled and verified by experiments to be suitable for laser ultrasound. Results indicate that the algorithm can effectively detect and locate small defects, and the image reconstruction speed is faster than that of the time domain SAFT algorithm. It can provide faster real-time technical solutions for laser ultrasonic NDT&E.
Methods The propagation mechanism of laser ultrasound in high-strength steel samples, as well as the interaction of the high-strength steel with circular through-hole defects, was numerically simulated by Comsol Multiphysics. The PSM-SAFT algorithm for laser ultrasound based on the principle of explosive reflection was then derived and adapted. Using the model data obtained by the finite element method, samples were imaged using longitudinal wave and shear wave signals, respectively. Then, the defect position error and signal-to-noise ratio were used for analyzing the imaging results of the two modes. Subsequently, in order to verify the applicability of the algorithm in actual application, an experimental device for laser excitation and Doppler-based vibrometer detection was built. PSM imaging was performed using one-dimensional scanning experimental data. Finally, different frequency ranges were selected by analyzing the center frequency of the selected acoustic waves, and the signal-to-noise ratio and imaging speed of the PSM imaging were compared.
Results and Discussions In the finite element simulation, we compared the time domain B-scan results of simulated data with and without circular defects. It can be found that there are obvious defect reflection signals in the defective B-scan graph (Fig. 3). If a difference between the model data with and without circular defects is made, the reflected signal of defects can be highlighted (Fig. 4). Then shear wave and longitudinal wave were used to perform PSM imaging on all simulation data in the frequency range of 50kHz--20MHz (Fig. 6). By comparing the signal-to-noise ratio of images and the position error of the defect, it is concluded that the calculation result of the shear wave is better than that of longitudinal wave. In the experiment, by analyzing the frequency spectrum of the time domain signal, it can be obtained that the center frequency of the shear wave is about 10MHz. Therefore, nine different frequency ranges are selected for PSM imaging, and the signal-to-noise ratio and imaging time of each image are compared (Fig. 9). It can be found that when the frequency range is 3.1--16.9MHz (90 frequency components in total), the PSM imaging results are optimized (Fig. 10). In addition, after comparing the imaging time lengths of the time domain DAS-SAFT algorithm and the reported frequency domain PSM algorithm, it is concluded that by using the same model, the calculation time of the frequency domain PSM algorithm is only 1/535--1/180 of that of the frequency domain DAS algorithm.
Conclusions In this paper, a frequency domain SAFT phase shift algorithm suitable for laser ultrasound is developed. In the process of PSM image reconstruction using experimental data, the broadband characteristics of laser-excited ultrasound are effectively used. By analyzing the center frequency of the acoustic wave mode used, a suitable frequency range is selected. This method not only avoids the influences of low-frequency noise and high-frequency noise, but also improves the signal-to-noise ratio of the image, and greatly reduces the consumed time for calculation. The results obtained by PSM algorithm for finite element simulation data and experimental data can accurately locate the defect position, which proves the feasibility of the algorithm in the field of laser ultrasound imaging.
