Beyond energy density: a regime-aware predictive framework and the identification of the flood mode in laser directed energy deposition of 316L stainless steel

The predictive potential of laser-powder directed energy deposition (LP-DED) remains limited because current models rely on oversimplified integrated energy metrics that cannot represent the process’s inherently complex, coupled physics. This study investigates the independent effects of laser power, beam diameter, and deposition speed to reassess these common metrics. The results indicate a clear limitation of linear (LED) and specific (SED) energy density metrics, as they fail to keep process outcomes consistent when parameters are changed independently. This failure occurs because these metrics cannot capture the nonlinear effects of process parameters, melting mode transitions, dynamic powder-melt-plume interactions affecting track geometry, or local thermal histories determining microstructure. Consequently, a regime‑specific modeling approach is required. Therefore, this study introduces a physics‑aware predictive framework developed for the stable conduction mode. This framework establishes power-law scaling models (R 2 > 0.84), validated using both our experimental data and independent datasets. Additionally, this investigation identifies a third mode termed the Flood mode. This mass-transfer-limited regime occurs at high powder feed per unit length, where the powder feed overwhelms the melt pool’s ability to absorb material. This finding suggests that the established LP-DED process map may need further refinement to include mass-transfer limits. Thus, reliable process prediction requires moving beyond common energy metrics toward regime-specific models that account for the independent contributions of each process parameter—an approach that substantially reduces experimental cost.