Detection and quantitative evaluation of surface defects in wire and arc additive manufacturing based on 3D point cloud

Abstract

ABSTRACT Wire and Arc Additive Manufacturing (WAAM) with high efficiency and low-cost is an economical choice for the rapid fabrication of medium-to-large-sized metallic components and has attracted great attention from scholars and entrepreneurs in recent years. However, defects such as porosity, and humps, could occur occasionally after each layer of deposition on weld bead surfaces due to disturbances and process abnormities. Detection and quantitative evaluation of weld bead defects is crucial to ensure successful deposition and the quality of the entire component. In this paper, a novel defect detection and evaluation system was developed for WAAM utilizing machine vision technology. The system incorporated new defect detection algorithms based on analysing the 2D curvature of the weld bead height curve and the 3D curvature of the weld bead point cloud. Furthermore, a defect evaluation algorithm was developed based on reconstructing the normal weld bead contour using geometric features extracted from the accumulated point cloud. This system enables the automatic detection of weld bead morphology during the WAAM process, offering important information about the location, type, and volume of defects for effective interlayer repairs and enhanced part quality.