Numerical modeling and optimization of process parameters in continuous extrusion process through response surface methodology, artificial neural network & genetic algorithm

The product of high complex profile, high strength, high productivity and excellent material properties with infinite length can be produced by Continuous Extrusion (CE) process. The numerical simulation of Aluminum (AA 1100) feedstock material at different wheel velocities, product diameter, feedstock temperature, die temperature and friction condition has been carried out using 3D simulation tool Design Environment for Forming (DEFORM-3D) in this paper. The development of mathematical model is carried out to investigate the influence of wheel velocity, extrusion ratio, feedstock temperature, die temperature and friction conditions on total load required for the deformation and extrusion of feedstock material through Response Surface Methodology (RSM). The statistical significance of mathematical model is verified through analysis of variance (ANOVA). The most optimum value of extrusion load has been found to be 136.4[Formula: see text]kN through iterative process of Genetic Algorithm (GA) using Artificial Neural Network (ANN). The optimized value of input process variables for minimum value of extrusion load obtained has been found to be 13 Revolutions per Minute (RPM) as wheel velocity, 5[Formula: see text]mm as product diameter, 0.95 as friction condition, 650[Formula: see text]C as feedstock temperature and 550[Formula: see text]C as die temperature. This paper with proposed methodology will be helpful for industries working in the area of CE in terms of minimizing energy consumption during production process of bus bars, tubes, wires, cables, sheets, plates, strips, etc.