Machine Learning Based Prediction of Energy Density for Additively Manufactured AlSi10Mg Samples by Laser Powder Bed Fusion Process

The laser Powder Bed Fusion (LPBF) process has recently become popular due to its ability to manufacture complex shapes with high precision. Among the various alloys for LPBF, AlSi10Mg stands out differently due to its lightweight nature, corrosion resistance, and design flexibility. The two important quality indicators for LPBF manufactured parts are surface roughness and Energy density, which should be within optimal range to ensure high-quality production. This study aims to improve this optimization process through a new approach—using a machine-learning model to predict the requisite laser energy density to manufacture AlSi10Mg. The study utilized machine learning models that were trained using four input parameters: Laser Power (W) Scan Speed (mm/s) Hatch Spacing (microns) and Layer Thickness (microns). Four different machine learning models were employed to predict the required Energy Density: Linear regression, Poisson regression XGBoost and ExtraTrees regression. The Poisson regression model yielded the results with an RMSE value of 11.13 and an R square value of 0.99. The XGBoost gradient boosting model also achieved outcomes with an RMSE of 52.18 and R values of 0.94. Overall, the trained machine learning models proved effective, in predicting the optimal energy density necessary for LPBF thereby enhancing manufacturing process accuracy and quality.