Process-informed micromechanical modelling of anisotropy and defect sensitivity in additively manufactured 316L stainless steel

Laser powder bed fusion (PBF-LB) of 316L stainless steel produces melt-pool-driven columnar grain structures that result in significant mechanical anisotropy. This study develops a predictive, process-informed 3D geometrically based process-structure (GBPS) framework integrated with crystal plasticity finite element (CPFE) modelling to simulate the tensile response of AM 316L. The framework replaces the need for experimental microstructure data with an idealised, computationally efficient 3D unit cell. Methodological optimisations, including the use of cuboid elements and regularised grain boundaries, achieve a 10-fold reduction in CPU time with negligible (<1%) loss in accuracy. The model successfully predicts the characteristic anisotropy of AM 316L, yielding ≈9% higher yield strength and ≈12% higher UTS in the transverse direction compared to the build direction. Furthermore, a defect-sensitivity study reveals that horizontal build orientations are significantly more susceptible to porosity-induced strain localisation, showing a ≈15% drop in ductility at 2% porosity. This framework provides a physically motivated yet lean tool for exploring process-structure–property relationships in PBF-LB alloys across varied scanning strategies.