Skip to content
Conference

3D CNN-Enabled Volumetric Phased-Array Ultrasonic Intelligence for Reliable Weld Defect Detection in Smart Electronic Inspection Systems

Oct 2026 · International Conference on Communication and Electronics Systems · pp. 684-689 · 0 citations · 25 references

Abstract

Reliability of the product depends on the inspection assessment in the industries that assess the manufactured Weld product. The Welding quality is detected by destructive and non-destructive testing, where the non-destructive testing (NDT) gains more interest due to destruction less destructive testing processes, and it is suitable for testing dense materials. In the proposed work, the Phased Array Ultrasonic dataset is used to find the defects in the welded material using the 3D CNN architecture. This work determines the depth wise extraction of features by the partitioning of the PAUT dataset. This performs the 3D volumetric convolution that helps to extract the features in depth of all the layers of the images, which provides the defect detection in an efficient way. The proposed work showed an accuracy of 95.89% with an F1 score of 96.65%. It is also seen that the proposed work attains FPR 0f 4.91%and FNR of 3.65% that improves the reliability of the 3D extraction processes for the PAUT dataset.

View source

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.