Intelligent Optimization of FDM Process Parameters for PLA/MWCNT Composites Using Taguchi and Random Forest Models
Abstract
FDM (Fused Deposition Modelling) has become one of the most popular technologies of additive manufacturing, which allows producing lightweight parts made of polymers with unique shapes and internal structure. At the same time, the mechanical characteristics of FDM parts largely depend on the material composition and the parameters of 3D printing, thus optimizing the tensile strength becomes an important task. The aim of this study is to optimize the tensile characteristics of polylactic acid (PLA) strengthened by multi-walled carbon nanotubes (MW-CNT) by exploring the influence of infill density, layer height and nozzle temperature while using the same Tri-Hexagon infill. Three variants of composite compositions containing 0.4 wt.%, 0.8 wt.% and 1.2 wt.% MW-CNT were considered to find the optimal printing parameters. An experimental plan based on the Taguchi L9 orthogonal array was used in order to minimize the number of tests and systematically analyze the influence of three 3D printing parameters with three levels each. In total, 27 tensile samples were produced using an Anycubic Kobra 2 Neo FDM printer and tested using ASTM D638 tensile testing standard. The obtained experimental data were analyzed using signal-to-noise ratio, regression, analysis of variance and Random Forest machine learning for tensile strength. The findings confirmed that the highest value of tensile strength of 29.95 MPa was obtained using the PLA+0.8 wt.% MW-CNT composite with 90% infill density, layer height of 0.18 mm and temperature of 210°C. The results from analysis of variance showed that infill density is the most important factor and contributed to 88.18-92.19% variations in tensile strength for all types of materials. In addition, the Random Forest model had the best prediction performance with test R² of 0.595. This research has confirmed that optimizing infill density is the best method for improving the tensile properties of PLA/MW-CNT composites.