Design–Modeling–Optimization workflow for part quality in DED-LB 304L stainless steel plates: Microstructure and mechanical properties

Laser-directed energy deposition (DED-LB) enables large-scale builds but poses challenges in geometric control and porosity. We developed and experimentally validated a Design–Modeling–Optimization workflow that maximizes relative density and targets the unit width to control part quality in DED-LB 304L stainless steel. From this workflow, width-indexed parameter trends Xi*(Wa,n) within the explored deposition strategy have been derived. First, a total of 136 runs (32/52/52) and observations were performed for the OO (one-layer, one-bead), OM (one-layer, multi-bead), and MO (multi-layer, one-bead) series by the Response Surface Methodology (RSM) design; then multiple linear regression (MLR) modeling was conducted, and models show high adequacy: Y1 follows the identity line within ±0.4 %, and Y2–Y3 are predominantly within ±7 %. After multiple-response optimization based on the desirability function (DF), trends Xi*(Wa,n) were provided for practical use. The trends indicate that larger target unit widths require higher optimal line energy (El* = P*/TS*). Using selected optimal parameter sets, three thin plates were fabricated and characterized in terms of geometry, microstructure, and mechanical properties. Three plates attain yield strengths 51–87 % above conventional 304L, while maintaining ductility consistent with Eurocode 3 requirements. Overall, the workflow controls part-quality, supplies parameter guidance through Xi*(Wa,n), and produces high-performance DED-LB 304L within the studied process window.