Predicting Fatigue Life in Additively Manufactured 316L Stainless Steel: An Integration of Machine Learning

ABSTRACT Additive manufacturing (AM) enables the fabrication of complex 316L stainless steel (SS) components for applications subjected to cyclic loading, making reliable fatigue life prediction essential. This study proposes a process‐independent machine learning framework for predicting the fatigue life of additively manufactured 316L SS using only tensile mechanical properties and fatigue loading conditions, without requiring detailed AM process or post‐processing parameters. A literature‐derived dataset covering multiple AM techniques was used to train and compare several models, with the multilayer perceptron showing the best predictive performance. Explainable AI was then employed to identify the most influential variables and interpret learned feature interactions. The results show that applied stress, elongation, and strength‐related properties are the dominant factors governing fatigue life, consistent with established fatigue mechanisms. The proposed method demonstrates the feasibility of transferable, property‐based fatigue life prediction across various AM processes, while reducing dependence on detailed processing data and extensive fatigue testing.