DOI: 10.1115/1.4072230 ISSN: 1087-1357

Improving the As-built Mechanical Properties in Laser Powder Bed Fusion Additive Manufacturing of Inconel 718 through Feedforward Control of Thermal History

Kaustubh Deshmukh, Mihir Darji, Essa Al Amiri, Harold (Scott) Halliday, Christopher B. Williams, Antonios Kontsos, Prahalada Rao

Abstract

This work investigates the laser powder bed fusion (LPBF) additive manufacturing of Inconel 718 and proposes a physics-guided feedforward control strategy to improve as-built mechanical properties by regulating the thermal history. Conventional LPBF parameter optimization relies on costly build-and-test studies and typically employs constant process parameters optimized for simple geometries. However, constant parameters can lead to non-uniform thermal gradients and cooling rates, resulting in microstructural variations and degraded mechanical performance. To address this challenge, a feedforward control framework was developed to maintain a near-constant cooling time (tc, [s]) throughout the build by adjusting laser power (P, [W]) on a layer-by-layer basis. The cooling time is nominally the inverse of the cooling rate. The required laser power modifications were determined prior to manufacturing using a rapid, experimentally validated thermal model. Inconel 718 ASTM E8/E8M tensile specimens were fabricated under two conditions: (i) constant empirically optimized process parameters and (ii) model-derived laser power adjustments designed to maintain a constant tc. Specimens manufactured under constant tc exhibited a more refined solidified microstructure than those produced with constant process parameters. The resulting microstructural refinement increased the mean as-built yield strength and ultimate tensile strength by approximately 5%. These results demonstrate that controlling thermal history through physics-based feedforward process control can improve mechanical performance without extensive empirical optimization. Although validated here using standardized tensile specimens, the proposed framework is generalizable to complex geometries owing to its first-principles physics-based foundation.

More from our Archive