Modeling and Experimental Investigation of Springback in Brass Alloy Sheet Metal V-Bending

One of the most important criteria in a good bending is springback. Springback can be defined as the amount of shape recovered after bending load. Springback value is somehow influenced by some factors like bending force, die opening, bend radius, bend angle, sheet metal thickness and punch holding time, and heat treatment. In this study, springback is investigated in v-bending for yellow brass material (UNS C27000 brass). The investigation of the springback will consider three parameters. The selected parameters to be varying will be the thickness of sheet metal (1.5mm 3mm and 5mm), the die opening (22 and 35mm), and the punch holding time (0 5 and 10 seconds) and all the other parameters will remain constant like the die angle and punch angle (85°). A full factorial multilevel design of experiment (DOE) method is used to conduct the experimental test. Finite element simulation (FEA) using ANSYS IMPLICIT, and experimental investigation results are compared. In addition, machine learning with ANN and SVM methods are used to predict the springback. The analysis of the result done with ANOVA showed interesting facts. The Pareto chart reveals that in comparison with the two other parameters, the sheet thickness has the highest significant effect on the springback and the die opening has the lowest significant effect. In overall the metal behaved approximately like the other metals because the springback decreased when the die opening increased and also when the holding time increased, and also when the sheet thickness increased. However, the brass material went in contradiction with the previous researches on the 3mm sheet thickness. Keywords: springback, finite element analysis (FEA), SVM, ANN, ANOVA.