Robust multi-objective optimization of L-PBF process parameters for AISI 316 L using NSGA-II, TOPSIS, and Monte Carlo simulation

This study optimizes the Laser Powder Bed Fusion (L-PBF) parameters for AISI 316 L stainless steel using a Central Composite Design (CCD) and Response Surface Methodology (RSM) to evaluate the effects of laser power, scanning speed, and hatching strategy on tensile mechanical responses. The scanning strategy is treated as an explicit categorical factor within the response surfaces, which allows its effect to be tested by ANOVA on the same footing as the continuous factors. ANOVA confirmed laser power and scanning speed as the most critical non-linear factors, with the optimized 0/90 strategy maximizing stiffness and strength. The predictive models exhibited high adequacy ( R²adj > 93% , non-significant lack-of-fit, and absolute prediction errors below 3.1% across the design space). Multi-objective optimization via NSGA-II and TOPSIS converged to a highly stable processing window: laser power between 168.8 W and 169.7 W, scanning speed of 717.2 mm/s, and the optimized 0/90 strategy. This optimal condition predicts an elastic modulus of 174.2 GPa, a yield strength of 498.9 MPa, and a strength coefficient of 673.2 MPa, representing gains of up to 35.0% over the poorest experimental condition. Experimental validation yielded relative errors below 8.3%, and Monte Carlo simulations ( $$n = 5000$$ ) confirmed substantial robustness with coefficients of variation strictly under 0.16%.