Effects of Printing Parameters on the Tensile Performance and Prediction of PLA-Based 3D Printed Objects

Abstract To examine the influence of key process parameters in fused deposition modelling on the tensile behaviour of Polylactic Acid (PLA) fabricated via three-dimensional printing. The influence of process parameters such as Raster Angle, Layer thickness, and Infill density was investigated to identify the best process combinations that enhance tensile behaviour, and machine learning (ML) models were employed to identify the most accurate predictive model for tensile strength and to establish an integrated framework for process optimisation. A total of 27 samples were fabricated using Fused Deposition Modelling with varying layer thicknesses (0.18, 0.22, and 0.25 mm) and infill densities (25%, 50%, and 75%). Tensile testing was conducted in accordance with ASTM D638 standards, followed by an analysis of variance (ANOVA) to quantify the significance of each process parameter. Five machine learning models, namely Decision Tree, Linear Regression, Random Forest Regressor, Gradient Boosting Regressor, and Support Vector Regression, were developed and evaluated to predict the tensile strength. The tensile strength of the fabricated specimens ranged from 23.91 MPa to 28.798 MPa. The highest tensile strength was achieved with a layer thickness of 0.18 mm, infill density of 75%, and a (45°, 90°) raster angle. ANOVA statistical analysis revealed that layer thickness was the most influential parameter on tensile strength, followed by infill density and raster angle in order of significance. Among the five ML models evaluated, the Decision Tree model demonstrated the highest predictive accuracy (Mean R² = 0.8577), outperforming the Linear Regression, Random Forest Regressor, Gradient Boosting Regressor, and Support Vector Regression. This study offers a comprehensive assessment of the effects of key FDM process parameters by integrating tensile testing, ANOVA, and machine learning within a unified framework. The proposed methodology not only identifies the optimum process parameter combination but also determines the most suitable predictive model for tensile strength, providing an effective framework for process optimisation and reducing the experimental effort in additive manufacturing.