Process Parameter Optimization for 3D Printing Using the Taguchi Method

Abstract

Although fused deposition modeling (FDM) is the most widely used material-extrusion AM process, the mechanical performance and productivity of parts produced by FDM are sensitive to process parameter settings which are often selected by trial and error. In this paper, the Taguchi method was used to optimize four FDM process parameters (layer height, infill density, nozzle temperature, and print speed) with three process levels for two competing responses ( tensile strength and print time). A Taguchi L9(3⁴) orthogonal array is used to decrease the number of required experiments from 81 (full factorial) to 9 trials (replicated by 3, 27 specimens) for generic polylactic acid (PLA). The infill density factor was identified as the most dominant factor in the signal-to-noise (S/N) ratio analysis and analysis of variance (ANOVA) for the replicated data with regard to tensile strength (68.99% contribution, significant at 1%) and layer height was identified as the most dominant factor in the signal-to-noise (S/N) ratio analysis and analysis of variance (ANOVA) for the replicated data with regard to print time (51.77% contribution, significant at 1%). The individual optima are 0.10 mm layer height, 75% infill, 220°C nozzle temperature, and 40 mm/s print speed, both for the maximum strength (predicted 49.02 MPa) and minimum print time (predicted 8.69 min), which means that it conflicts on three of the four parameters. This tradeoff was successfully overcome by the equal weighting grey relational analysis (GRA) which determined a compromise setting of 0.30 mm layer height, 75% infill, 220°C nozzle temperature, and 80 mm/s print speed; the print speed (47.89%) and the nozzle temperature (30.34%) had the greatest contribution to the compromise grade. The single-response optimum and the GRA compromise are not exactly the same as one of the nine physically tested trials and thus confirmation experiments at the recommended setting are proposed before the results are adopted for production application. A resource-saving methodology and example for optimizing