Simulation and optimization of hot extrusion process to produce rectangular waveguides

In this study, the Taguchi method and genetic algorithm are employed to optimize the die geometry in order to obtain the minimum magnitude of extrusion force to produce the rectangular waveguides. Waveguides are hollow metal tubes for transporting electromagnetic energy from one region to another. A combination of process parameters was selected using orthogonal array Taguchi method. The results were then simulated by finite element method (FEM). By using the analysis of variance (ANOVA) method, the effective parameters and their interactions were determined. By using the FEM results, an artificial neural network (ANN) model was trained, and the effects of die factors on the extrusion force were studied for constant thermal and frictional conditions. Moreover, the optimum conditions were predicted by means of genetic algorithm (GA) method which was verified by experimental procedure. The optimum die length and billet hole diameter were obtained at L = 20 and D = 3 cm, respectively, for the initial billet temperature of 450 °C and friction coefficient of 0.1.