Defect Modeling During the SLM Process for Manufacturing Microwave Devices

This paper intends to address the issue of crack formation and other flaws during the selective laser melting (SLM) additive manufacturing process. To achieve this objective, image processing, 3D modeling, and deep-learning techniques are employed to generate a 3D defect model, while data statistics are utilized for enhancing and optimizing the entire additive manufacturing process, including adjusting manufacturing process parameters, optimizing strategies, reducing defects, and improving the yield rate of SLM. After training and adjustment, the crack recognition accuracy of the final model can reach 92.3%.