Parametric Prediction of FDM Process to Improve Tensile Properties Using Taguchi Method and Artificial Neural Network

Abstract

Fused deposition modeling (FDM) is a popular 3D printing technique that creates parts by heating, extruding, and depositing filaments made of thermoplastic polymers. The processing parameters have a considerable impact on the characteristics of FDM-produced parts. This paper focuses on the parametric prediction of the FDM process to predict ultimate tensile strength and determine a mathematical model using the Taguchi method and Artificial Neural Network. Five manufacturing variables, such as layer thickness, print speed, orientation angle, number of parameters, and nozzle temperature at five levels, are used to study the mechanical properties of PLA material to manufacture specimens using FDM 3D printer. The specimens are produced for tensile tests in accordance with ASTM-D638 standards, and the process parameters are established using the Taguchi orthogonal array experimental design technique. The results proved that the printing process parameters significantly impacted the tensile strength by changing the tensile test values between 37 MPa and 53 MPa. Also, the neural network predicted the tensile strength values, and the maximum error was equal to 8.91%, while the mathematical model had a maximum error equal to 19.96%.