A deep learning-based model for defect reorganisation in welding/wire arc additive manufacturing

Welding and Wire Arc Additive Manufacturing (WAAM) are crucial processes in various industries, but defects can significantly impact the quality and integrity of the final product. The study proposes a deep learning-based model for automated defect recognition in WAAM. The primary objective is to develop a robust and accurate system that can identify welding defects using Convolutional Neural Networks (CNN), a deep learning model. The research methodology involves inputting weld bead images into the model, which is then trained to recognise different defects. During the training phase, images depicting various defects were used to train the DL model, enabling it to learn the distinctive features and patterns associated with each type of defect. The trained model was later tested, and it demonstrated an accuracy rate of 88% in classifying welding defects, showcasing its potential for real-time monitoring and quality control in welding/WAAM applications. The findings of this study contribute to the advancement of intelligent manufacturing systems and highlight the effectiveness of deep learning techniques in enhancing the reliability and efficiency of welding processes.