A hybrid approach based on numerical, statistical and intelligent techniques for optimization of tube drawing process to produced squared section from round tube

Abstract

In the tube drawing process, there are a bunch of parameters which play key role in process performance. Thus, finding the optimized parameters is a controversial issue. Current study aimed to produce a squared section of round tube by tube sinking process. To simulate the process finite element method (FEM) was used. Then, to find a meaningful kinship between process input and output parameters the developed FE model was associated with the design of experiment based response surface methodology (RSM). The sufficiency of each model was checked by analysis of variances. Further, the SA (simulated annealing) was associated with RSM models to find the optimal solution regarding maximum thickness distributions and minimum force and dimensional error. Hereafter, for performing accurate optimization, the principal component analysis was used to find the appropriate weight factor of each response. The obtained results were in right agreement with those derived from simulation and confirmatory experiment.