A statistical model for estimation of bending angle in laser bending of two-layer steel-aluminum sheets

Abstract

Laser bending of sheet metals can be considered as an economic and flexible forming method since it does not need a specific forming die. The incorporation of this method in bending two-layer sheets adds to the complexity of the bending process. The bending angle obtained in this method depends on several influential parameters. In bending two-layer sheets, the accuracy and controllability of this method are increased by the determination of the effects of process parameters on the bending angle. It seems to be necessary to derive a relationship between the influential parameters and their effects on the bending angle of the two-layer sheets. In this paper, the parameters affecting the bending angle of two-layer steel-aluminum sheets have been investigated through conducting finite element analysis and experimental tests. Three-dimensional simulations were performed in Abaqus software using coupled temperature-displacement analysis. The experimental tests were designed and performed to validate the numerical results. Comparing the numerical and experimental results, the maximum error in the prediction of bending angle was obtained as 10.3 %. After ensuring the accuracy of the simulation results, the experiments were designed using the response surface methodology to investigate the effects of the influential parameters on the bending angle, including the laser power, laser scan speed, and the number of passes of the laser on the specimen. The results of the ANOVA test showed that the laser power and the number of passes had the most and least effects on the bending angle, respectively. By increasing laser power from 500 W to 1500 W, the bending angle increased by 103 % and 63 % with the number of laser passes equal to 15 and 25, respectively. The simulation results showed that with the increase in power, the number of laser passes, and the decrease in laser scanning speed, the changes in the bending angle increase. This issue is provided that the necessary temperature gradient remains stable in the sheet. Power, laser scanning speed, the product of power in laser scanning speed, number of laser passes, the product of power in the number of passes, and product of the number of passes in laser scanning speed are respectively the most influential parameters on bending angle. Finally, a mathematical relationship was proposed based on the regression method to predict the bending angle based on the significant process parameters and their interactions.