A hybrid ANN/PSO optimization of material composition and process parameters for enhancement of mechanical characteristics of 3D-printed sample

Abstract

Purpose

This paper aims to study the effects of inorganic CaCO₃ nanoadditives in the polylactic acid (PLA) matrix and fused filament fabrication (FFF) process parameters on the mechanical characteristics of 3D-printed components.

Design/methodology/approach

The PLA filaments containing different levels of CaCO₃ nanoparticles have been produced by mix-blending/extrusion process and were used to fabricate tensile and three-point bending test samples in FFF process under various sets of printing speed (PS), layer thickness (LT), filling ratio (FR) and printing pattern (PP) under a Taguchi L27 orthogonal array design. The quantified values of mechanical characteristics of 3D-printed samples in the uniaxial and the three-point bending experiments were modeled and optimized using a hybrid neural network/particle swarm optimization algorithm. The results of this hybrid scheme were used to specify the FFF process parameters and the concentration of nanoadditive in the matrix that result in the maximum mechanical properties of fabricated samples, individually and also in an accumulative response scheme. Diffraction scanning calorimetry (DSC) tests were conducted on a number of samples and the results were used to interpret the variations observed in the response variables of fabricated components against the FFF parameters and concentration of CaCO₃ nanoadditives.

Findings

The results of optimization in an accumulative scheme showed that the samples of linear PP, fabricated at high PS, low LT and at 100% FR, while containing 0.64% of CaCO₃ nanoadditives in the matrix, would possess the highest mechanical characteristics of 3D-printed PLA components.

Originality/value

FFF is a widely accepted additive manufacturing technique in production of different samples, from prototypes to the final products, in various sectors of industry. The incorporation of chopped fibers and nanoparticles has been introduced recently in a few articles to improve the mechanical characteristics of produced components in FFF technique. However, the effectiveness of such practice is strongly dependent on the extrusion parameters and composition of polymer matrix.

Citation

Seyedzavvar, M. (2023), "A hybrid ANN/PSO optimization of material composition and process parameters for enhancement of mechanical characteristics of 3D-printed sample", Rapid Prototyping Journal, Vol. 29 No. 6, pp. 1270-1288. https://doi.org/10.1108/RPJ-10-2022-0338