Fatigue Crack Growth Evaluation in SLM C300 Margining Steel Using Data‐Driven Predictive Models

ABSTRACT Selective laser melting (SLM) enables the fabrication of complex metal components but poses challenges due to fatigue failure under cyclic loading. This study investigates fatigue crack growth (FCG) in SLM C300 margining steel using compact tension (CT) specimens built in X , Y , and Z directions, tested per ASTM E647 with a constant load ratio ( R = 0.1). Monotonic tests revealed direction‐dependent yield stress, with X being highest and Y lowest. To predict fatigue crack growth rate (FCGR) and stress intensity factor (SIF), five nonlinear regression models were evaluated. Given the nonnormal distribution of FCGR and SIF, the models were refined using load ratio ( R = P min / P max ) and stress intensity factor ratio ( R k = ( ΔK ) min /( ΔK ) max ). Locally weighted regression (LWR) achieved the highest accuracy, followed by polynomial and hyperbolic tangent regression. Logistic regression improved at higher SIF values, whereas power regression, despite aligning with Paris' law, performed poorly. Model choice and material properties are crucial for reliable FCG prediction in SLM components.