Enhancing additive manufacturing quality and productivity using machine learning: a study on laser powder bed fusion of AISI 316L-2.5%Cu

Abstract Metal additive manufacturing (AM) has advanced the fabrication of complex metal components, providing remarkable precision and flexibility in producing different geometries. Integrating artificial intelligence (AI), particularly machine learning (ML), further improves AM by uncovering complex relationships within manufacturing processes and enabling precise quality control. In this study, ML is employed to optimize process parameters in laser powder bed fusion (L-PBF) for AISI 316L-2.5%Cu components, focusing on minimizing defect content and enhancing productivity. Seven supervised ML algorithms, including Bayesian Regression, Decision Tree Regression, Gradient Boosting Regression, Gaussian Process Regression, K-Nearest Neighbors Regression, Random Forest Regression, and Support Vector Regression (SVR), were evaluated for their predictive accuracy using relative density as a target metric. Among these, SVR demonstrated the highest accuracy with a Mean Absolute Error (MAE) of 0.601 and a coefficient of determination (R 2 ) of 0.842. The optimized process parameters—laser power of 200–250 W, scan speed of 800 mm/s, and hatch distance of 0.13 mm—achieved a relative density exceeding 99.5% while maintaining high productivity. The optimized parameters derived from this approach provide a robust framework for balancing quality, productivity, and defect minimization in AM processes.