Statistical Distributions of Morphologically Classified Defects in Metal Additive Manufacturing with Implications for Fatigue Life Prediction

Metal additive manufacturing (AM) enables the production of complex components but often introduces defects with varied morphology, size, and distribution—factors that significantly affect fatigue performance. This study investigates the role of statistical distributions in defect analysis and discusses their implications for probabilistic fatigue modeling methods. Metrics including the coefficient of determination, Kolmogorov–Smirnov goodness‐of‐fit test, and quantile–quantile plots are employed to evaluate the fit between proposed distributions and actual defect distributions. A novel triplewise comparison method is proposed to quantify overlap among morphological features and enhance defect classification. To assess the influence of distribution choice, a simulation study using synthetic datasets—generated with best‐fit and uniformly applied distributions—demonstrates that accurate distribution selection is critical for reliable feature ranking and defect differentiation. By leveraging these statistical methods, manufacturers can more effectively identify and mitigate critical defects, thereby improving the reliability and industrial applicability of AM technologies.