In this article, the contributions are to further study and verify a bulk-wave electromagnetic acoustic transducer (EMAT) for the transverse crack detection of rail base and to propose an auxiliary method for achieving efficient detection. There is no verification for rail treated as a plate in the previous publication. For the verification, the wedge-like rail base is reasonably simplified into the front view and the side view to replace the 3-D finite-element model (FEM) simulation. In addition, the experiment to detect an artificial transverse crack of the standard rail specimen is carried out to verify the treatment of the wedge-like rail base. Both simulation and experimental results demonstrate that the wedge-like rail base can be inspected as a steel plate using a spiral-coil EMAT, but the only time of flight is hard to detect the transverse crack of the wedge-like rail base due to low SNR. For solving the problem of low SNR, this article proposes an auxiliary method of watching for the disturbances of the back-wall echo. Also, the limit condition of the auxiliary method is analyzed for practical applications.

To further study the dominant mechanism, this paper utilizes the analytical model to analyze the spiral-coil electromagnetic acoustic transducer (EMAT) above a ferromagnetic plate. The analytical model includes the spiral coil as well as the sintered NdFeB (S-NdFeB) permanent magnet (PM) whose analytical solutions inside the ferromagnetic material are rarely presented relative to the coil. Because of this point, this paper presents the analytical model for S-NdFeB PM by the equivalent ampere current to compute the magnetic flux density and compare the computational results with that by FEM and measurement. The comparison shows that the analytical model is valid to analyze the spiral-coil EMATs. Based on the analytical model, both the Lorenz force and the magnetostrictive force are computed and compared, as well. Finally, it is found that the Lorentz force is a dominant mechanism in spiral-coil EMATs, which is verified by the experimental result.

Deep learning (DL) has been frequently applied in the image reconstruction of electromagnetic tomography (EMT) in recent years. It offers the potential to achieve higher-quality image reconstruction. Among these, research on samples is relatively scarce. Samples are the cornerstone for both large and small models, which is easy to ignore. In this paper, a deep learning electromagnetic tomography (DL-EMT) model with nine elements is established. Complete simulation and experimental samples are obtained based on this model. On the sample sets, the reconstruction quality is observed by adjusting the size and configuration of the training set. The Mann–Whitney U test shows that beyond a certain point, the addition of more samples to the training data fed into the deep learning network does not result in an obvious improvement statistically in the quality of the reconstructed images. This paper proposes a CC-building method for optimizing a sample set. This method is based on the Pearson correlation coefficient calculation, aiming to establish a more effective sample base for DL-EMT image reconstruction. The statistical analysis shows that the CC-building method can significantly improve the image reconstruction effect in a small and moderate sample size. This method is also validated by experiments.

The three-dimensional (3D) reconstruction of Electromagnetic Tomography (EMT) is an important task for many applications, such as the non-destructive testing of inner defects in rail systems. Additionally, image reconstruction algorithms utilizing deep learning methods have been verified to be useful in recent years. Therefore, the interpretability of deep learning is a question that is relevant to its application in other areas. This paper proposes an innovative rotational convolution pattern, Conv-P, for convolutional neural network (CNN) image reconstruction in a 3D EMT system. This pattern is based on the projection relationships inherent in tomographic imaging, where each convolution is performed on adjacent projections along the excitation rotation direction. The advantage of this pattern is that it can generate the convolution process by utilizing the 3D structural information from real sensors. To verify the effectiveness of this convolution pattern, we constructed a 3D dual-layer 16-coil EMT model and tested its image reconstruction performance. The results demonstrate that, compared with two common convolution patterns, Conv-P achieves a 4.7% and 4.1% increase in the Image Correlation Coefficient (CC), a 19.8% and 13.1% reduction in the Relative Image Error (IE), a 0.67% and 1.59% increase in the Peak Signal-to-Noise Ratio (PSNR), and a 3.24% and 0.74% increase in the Structural Similarity Index Measure (SSIM).