Multi-physical prediction and regime identification of lack-of-fusion porosity in selective laser melting

Lack-of-fusion porosity is one of the most critical defects in selective laser melting, severely impairing the mechanical integrity of printed components. To address this issue, a three-dimensional computational fluid dynamic model based on the open-source OpenFOAM package is employed to investigate the formation mechanisms of lack-of-fusion porosity under varying laser powers, scanning speeds, and thermal conductivities of solid and liquid phases. The model incorporates multiple physical effects, including surface tension, Marangoni flow, recoil pressure, and laser reflection, enabling accurate simulation of melt pool dynamics. Moreover, two revised parameters, the revised linear energy density and revised heat diffusivity, are proposed to characterize laser-material interactions. Additionally, a novel criterion based on the time-integrated temperature above the melting point is introduced to accurately describe the final morphology. Our results reveal that melt pool penetration depth is a critical indicator of lack-of-fusion formation. Furthermore, nonlinear relationships such as M-shaped and wave-like boundaries are observed between laser and thermal indexes, governed by the interplay between heat conduction and evaporation. This study provides a quantitative framework, including computational fluid dynamic model, dimensionless parameter, and regime identification, to predict the lack-of-fusion porosity and offers guidance for optimizing selective laser melting processing parameters to achieve high quality and defect-free manufacturing. These simulation-guided, dimensionless regime maps are predictive within the investigated ranges and assumptions, supported by literature-based calibration, uncertainty quantification, cross-validated classification, and independent experimental validation.