Model-assisted ultrasonic calibration using intentionally embedded defects for in-process weld inspection

Automated in-process Non-Destructive Testing (NDT) systems are rapidly gaining traction within the manufacturing industry as they reduce time and costs. When considering calibration verification of such systems, creating defects known geometry nature during deposition a weld can: (I) help examine capability automated system to detect characterise defects, (II) be used form database signals associated with different defect types train intelligent classification algorithms, (III) act basis for gain inspection at high temperatures, where ultrasound beam can skewed result velocity gradients. In view this, this paper investigates two unique methodologies introducing: (a) lack fusion by embedding tungsten in (b) artificial cracks quenching imitate real cracking scenarios. According results Phased Array Ultrasound (PAUT) inspections, were successful. The validity inspections was also verified extracting micrographs from defective sections welds, model-based simulations carried out better understanding wave propagation path interaction generated defects.