Improve springback prediction through the development of empirical models and optimize springback for the air-bending sheet metal process for Mild Steel, Aluminum, and Stainless Steel

Abstract

Products or components manufactured by the sheet metal process are indispensable in this contemporary era, from daily-used metal jars to high-tech air vehicles. Nonetheless, there are a variety of sheet metal manufacturing processes for fabricating these goods; the air-bending sheet metal process is one of the most commonly conducted methods. After air bending, commonly used metals — Mild Steel, Aluminum, and Stainless Steel — tend to demonstrate an unwanted characteristic called springback, which requires being controlled to produce extremely precise parts. Therefore, this study aimed to develop an empirical model that would improve springback prediction which would ultimately assist in minimizing the springback effect in bending operations. That pragmatic model was supposed to be derived from four independent parameters, such as die gap, punch radius to sheet thickness ratio, set angle, and material. In addition to this, rigorous optimization of the input variables — springback was vehemently affected by those — was also a goal of this research. To achieve both objectives thoroughly, several approaches were incorporated sequentially: Box-Behnken experimental design, experimental data collection, conducting an ANOVA test, empirical model selection, model development, confidence interval check, model validation, and finally numerical optimization. Design Expert^® software, Universal Testing Machine, and Bevel Protector were utilized for pursuing this study’s methodology. Thenceforth, the linear model was chosen in light of statistical analysis. In the validation phase, it was found that the developed linear models were able to prognosis springback with acceptable errors. Moreover, the highest sensibility of springback for one process parameter — punch radius to sheet thickness ratio — was also revealed from the developed linear models. Another feature of this study’s analysis was to optimize the given factors for four different circumstances. It illustrated that different kinds of optimization were possible in light of boundary constraints. Overall, this study has some significant insights for sheet metal industries; cost, process, and design optimization; and a bright path for forthcoming research.