Abstract
The rapid development of metal additive manufacturing (AM) technology has led to its widespread application across various industries, with quality control of AM components being a current focal point of research. To enable on-line defect detection during the metal AM process using laser ultrasonic technology, this paper proposes a defect detection method based on sparse scanning. By employing sparse scanning to collect data, the method significantly improves detection efficiency while accurately characterizing the position and morphology of defects. This study thoroughly investigates the interaction mechanisms between ultrasonic waves and defects. Ellipses are drawn based on the propagation paths of ultrasonic waves and scanning parameters, and the defect edges are characterized by the intersection points of adjacent ellipses and their corresponding tangent points. Five sets of experiments were designed to examine four typical types of defects, and multiple ultrasonic signals were collected using sparse scanning. For defects larger than 1 mm, the experimental results demonstrate that the proposed method can effectively output the position and edge morphology of the defects. Compared to the Synthetic Aperture Focusing Technique (SAFT) method, this method requires only 15.5% of the scanning data, achieves 31.92% of SAFT's computational efficiency, and has a Mean Absolute Error (MAE) of 27%. For internal hole defect with a diameter of 0.4 mm, consistent results with the SAFT method were obtained using 32% of the data. These experiments validate that the proposed defect detection method based on sparse scanning is an efficient method suitable for on-line defect detection in additive manufacturing.