The article presents an innovative approach to enhancing ultrasonic non-destructive testing (NDT) by using an automated signal analysis. In the context of Industry 5.0, where humans work alongside robots and artificial intelligence (AI), ultrasonic NDT has become a vital tool for quality inspection in industrial processes. A key part of this process is signal processing, where traditionally relies on skilled professionals to interpret the data. However, systematic research on automating and optimizing this process for real-time, intelligent analysis remains limited.

The study addresses this gap by focusing on the time-of-flight (ToF) estimation, a critical metric in ultrasonic testing that measures the duration an ultrasonic pulse takes to travel to and from a defect in a material. The researchers conducted a comparative analysis of different ToF algorithms and introduced a novel method, the Defect Peaks Tracking Model (DPTM), for the automatic diagnosis of ultrasonic echo signals. Leveraging the Hilbert transform and wavelet denoising, this method aims to improve the accuracy of locating defects in materials such as steel plates and pipelines.

To validate the method, the research team designed and manufactured a mechanical fixture using 3D printing technology, facilitating more efficient data collection during pipeline inspections. The results demonstrated that the DPTM reduced the error in auto-diagnosis to just 0.25% for steel plates and 1.25% for pipelines, significantly outperforming traditional cross-correlation methods.

The real-time identification capability of the DPTM holds significant potential for future integration with AI technologies, such as machine learning and deep learning, to further enhance the automation and intelligence of industrial health inspections. This advancement could lead to more reliable and efficient methods for defect detection in materials, contributing to safer and more effective industrial operations.