Parameter development for near-fully dense LPBF-produced stainless steel 316 using a design of experiment approach

Abstract

This study aims to develop parameters that minimize porosity in SS316 using a Design of Experiments approach (DOE). Employing the D- Optimal (D-Optimal) method and ANOVA, we determined a quadratic model for predicting porosity. The model was validated through three-point predictions, demonstrating high accuracy. The study utilised the bootstrapping method to identify the minimum sample size required for reliable predictions, establishing that 15 samples are sufficient. This was verified by plotting three sets of 15 random samples, where actual vs. predicted values showed strong agreement. The study achieved the highest density of 7.94 g/cm³, corresponding to 99.6% relative density. These findings offer a robust framework for optimising material properties to reduce porosity effectively.