Application of artificial neural network to evaluation of dimensional accuracy of 3D‐printed polylactic acid parts

Abstract

Additive manufacturing (AM) has begun to replace traditional fabrication because of its advantages, such as easy manufacturing of parts with complex geometry, and mass production. The most important limitation of AM is that dimensional accuracy cannot be achieved in all parts. Dimensional accuracy is essential for high reliability, high performance, and useful final products. This study investigates the impact of printing parameters on the dimensional accuracy of samples fabricated through fused deposition modeling (FDM), an additive manufacturing (AM) method utilizing polylactic acid (PLA) material. The experimental design process was performed using Taguchi methodology. ANOVA was used to determine the most important parameter affecting accuracy. Based on experimental studies, the optimal printing parameters for parts are determined as follows: concentric infill pattern, 3 mm wall thickness, 70% infill density, and a layer thickness of 200 μm. Artificial neural network (ANN) was used in the evaluation and prediction of the results. The R‐square (R2) performance evaluation criterion was above 95% from the ANN results. This value shows that the results are significant. The data acquired from this study may assist in identifying optimal parameters that contribute to the fabrication of samples with high dimensional accuracy using the FDM method.

Insight into constitutive theories of 4D printed polymer materials: a review

“…For example, Yan et al [146] discovered a new thermoset SMP with high recovery stress and modest glass transition temperature using ML algorithms with a small dataset. ANN have also been used to predict dimensional errors during the 3D fabrication of SMPs [148,149] and glass transition temperatures [150].…”

Experimental study of the effect of different 3D printing parameters on tensile strength, using artificial neural network

“…It takes into account complex relationships between variables, which can be difficult to do analytically [18]. However, it is important to note that this requires significant data collection and machine learning expertise to develop and adjust the model properly [19]. In addition, the use of AI to optimize 3D printing parameters can greatly facilitate the process of adjusting the parameters to get the best performance.…”

Restoration of a wet corrosion-resistant composite filament for material extrusion process

“…Taguchi (Gray, 1988) developed a family of FFE matrices (an orthogonal array) that has been generally adopted to optimize the design parameters and significantly reduce the overall test time and experimental costs by following a systematic approach to limit the number of experiments and tests. Various studies have utilized Taguchi's technique for MEX process (Arockiam et al , 2024; Gunes et al , 2024; Kam et al , 2023; Li et al , 2024; Uludag and Ulkir, 2024).…”