Modeling and Optimization of 3D Printed PLA Material for Maximum Flexural Strength Using Multiple Nonlinear Neuro Regression Analysis

Abstract

Fused deposition modelling (FDM) is a growing additive manufacturing method to produce complex objects without geometrical limitations. In addition, mechanical strength, dimensional accuracy, product development cycle time, and surface properties can be improved depending on the application of the best settings of design variables. There are various printing parameters which influence the mechanical properties and quality of FDM parts. In this study, appropriate printing parameters were determined to obtain desired quality on mechanical properties and dimensional accuracy. Then full factorial design was employed to form experiment set including process parameters. Multiple nonlinear neuro-regression analysis was used for modeling of FDM process. The present study aims at optimization of the FDM process parameters including infill pattern, infill density and build orientation on flexural strength and strain for polylactide (PLA) material. In this regard, optimization algorithms Differential Evolution and Nelder Mead were used to find the best design or elite designs. Third-order polynomial model and hybrid model including polynomial and logarithmic terms were employed as an objective function to define physical phenomena regarding flexural strength and strain, respectively. It was found that (i) maximum flexural strength as 99.66 MPa using a cubic pattern, flat orientation, and 90 % infill density, (ii) minimum ultimate strain as 1.102 % for gyroid pattern, flat orientation, and 47 % infill density.