MACHINE LEARNING APPROACH FOR DEFECT PREDICTION IN METAL 3D PRINTING FOR AEROSPACE APPLICATIONS

Abstract

In the aerospace industry, additive manufacturing (AM) has revolutionized the production of lightweight, high-strength components, such as engine parts and structural elements. The ability to create intricate geometries and reduce material waste is particularly beneficial for aerospace applications, where performance and weight savings are paramount. However, ensuring the quality and reliability of these components remains a challenge, particularly in mass production, related to material quality, expensive processes, and longer computational times than conventional manufacturing methods. The optimization process called machine learning (ML) can decrease the influence of those limitations and make AM applicable to mass market production. This research proposes a solution in the form of a Decision Tree Classification Machine Learning Algorithm to predict the possibility of defect occurrence in additive manufacturing processes. The research aims to create a machine learning algorithm that provides a new pathway for printing defect-free parts without the expenses and time-consuming trial-and- error testing typically required in Powder Bed Fusion (PBF). Correspondingly, the defect susceptibility index was developed to eliminate defect formation before the part’s manufacturing process, and the hierarchical order of importance of mechanistic variables on defect formation was determined. The obtained results demonstrate that a trained machine-learning algorithm can print defect-free parts without incurring expenses and time-consuming trials. This approach not only enhances the reliability of additive manufacturing in aerospace applications but also paves the way for its broader adoption in mass production. By integrating CFD analysis, machine learning, and experimental validation, this research provides a comprehensive solution to the challenges faced in additive manufacturing. The proposed methodology ensures the production of high-quality, defect-free components, making additive manufacturing a viable option for the aerospace industry and beyond