Prediction and Optimization of Tensile Strength Values of 3D Printed PLA Components with RSM, ANOVA and ANN Analysis

Authors : Fuat Kartal, Arslan Kaptan

Pages : 51-60

Doi:10.36222/ejt.1561857

View : 69 | Download : 68

Publication Date : 2025-07-01

Article Type : Research Paper

Abstract :This study evaluates the comparative effectiveness of Response Surface Methodology (RSM), Analysis of Variance (ANOVA), and Artificial Neural Networks (ANN) in predicting and optimizing the tensile strength of 3D-printed PLA components. Key process parameters—including layer thickness, infill density, print speed, temperature, and build orientation—were systematically varied to analyze their impact on tensile strength. The results indicate that RSM and ANOVA offer higher prediction accuracy compared to ANN, with lower deviation rates (0.65%, 0.18%, and 3.43% for RSM; 0.20%, 0.12%, and 3.25% for ANOVA) versus ANN (5.93%, 3.88%, and 6.26%). The analysis revealed that layer thickness plays the most significant role in tensile strength, followed by temperature, infill density, build orientation, and print speed. The optimal combination of parameters—0.20 mm layer thickness, 50% infill density, 50 mm/s print speed, 220°C nozzle temperature, and 90° build orientation—yielded a maximum tensile strength of 55.506 MPa. These findings highlight the importance of parameter optimization in improving the mechanical properties of FDM-printed components. The study provides valuable insights for enhancing the reliability and efficiency of additive manufacturing processes, paving the way for future research on hybrid modeling techniques and alternative material applications.
Keywords : Response surface methodology, Analysis of variance, Tensile strength, Fused deposition modeling, 3D printing, PLA components, Artifical neural networks, Process optimization

ORIGINAL ARTICLE URL