Abstract
The rapid identification of process windows in laser powder bed fusion (L-PBF) additive manufacturing garnered significant attention for its ability to reduce upfront engineering costs. This study focuses on accelerating the development of process windows by targeting the elimination of specific-size pore defects in L-PBF IN718. A novel relative density-porosity similarity evaluation method (DPSEM) is introduced to evaluate the reliability of porosity data derived from computed tomography (CT). Using the response surface method, the fully dense forming window (e.g., relative density ≥ 99%) was accurately located within a wide process parameter range (18-1000 J/mm(3)) in a single test. Comparative analysis with the relative density (RD) model highlighted differences in solution set distribution, positioning efficiency, microstructure, and performance within the process window. Results demonstrate that the proposed method effectively eliminates specified size defects (90 μm), achieving a maximum density of 99.5% alongside excellent mechanical properties, including an ultimate tensile strength of 1155 MPa and a yield strength of 908 MPa. In contrast, the RD model achieved a lower maximum density of 98.5%, with mechanical performance compromised by significant MC compound precipitation and keyhole pore accumulation, resulting in an ultimate tensile strength slightly exceeding 910 MPa.