Grain Scattering Noise Modeling and Its Use in the Detection and Characterization of Defects Using Ultrasonic Arrays

Design and Application of Partial Immersion Focused Ultrasonic Transducers for Austenitic Weld Inspection

Austenitic stainless steel is a widely used material in the industry, and the welding technique enables stainless steel components to have different shapes for different applications. Any flaws in the weld will degrade the performance of the austenitic component; thus, it is essential to ultrasonically and nondestructively test flaws in welds to ensure service safety. Recently, weld inspection has been performed using contact transducers, but missed detections or false positives for flaws in welds usually occur due to a poor coupling condition in the detection, a low signal-to-noise ratio, and instantaneous noises. In this study, a partial immersion focused (PIF) ultrasonic transducer is designed and used for austenitic weld inspection to address the above issues. The detailed design and manufacture of the PIF transducer are described, and the advantages of the transducer are shown by comparing the results detected using different kinds of transducers. In addition, in order to suppress false positives, a B-image method optimized using a time-dependent threshold is proposed. Experiments are performed to detect flaws in a welded specimen. All the artificial flaws are evaluated using the developed transducer and the proposed method, but minor flaws are mis-detected when planar transducers are used, verifying the method proposed in this paper.

Subwavelength ultrasonic imaging using a deep convolutional neural network trained on structural noise

Subwavelength ultrasonic imaging (SUI) can detect subwavelength flaws beyond the diffraction limit, however, SUI sometimes fails to clearly reveal flaws in C-scans when the signal-to-noise ratio (SNR) is low. In this work, a convolutional neural network (CNN) that takes structural noise into account is developed for SUI to distinguish flaw echoes from structural noise. The network contains a regression CNN for learning features from the structural noise and a learnable soft thresholding layer for classification. Experiments show that the proposed method performs well for imaging subwavelength flaws at different depths and of different sizes. It achieved an F₁ score of 97.69 ± 1.56% in detecting flaws as compared to the enhanced ultrasonic flaw detection method with time-dependent threshold. As an example of general application of the method, we also performed SUI on natural flaws in a spheroidal graphite cast iron specimen. The results show that the method can achieve SUI without a theoretical backscattering model and is not limited by noise distribution, multiple scattering, or complex microstructures. Furthermore, the network does not need to prepare flaw echoes for training.