The Prediction Of Subsequent Formality Of Sheet Metal After Passing Semi-Circle Drawbead

Drawbead is an important control way in the progress of sheet metal forming. As the development of modern industry and economy, it seems much more important to provide the theoretic support for the design of drawbead. In the past investigation, people focused on the mechanical properties of the drawbead, but after the sheet through the drawbead, the forming performance will change a lot, and it will have a greater influence to the subsequent forming of sheet metal. Therefore, this paper carries out a systematic study on the semi-circular drawbead of sheet metal subsequent formality.In the paper, the drawbead simulation test device is used to do experiments that three kinds of sheet metals called BUFDE+Z, DC54D+ZF and ST15 pass through the semi-circular drawbead. In the different structural parameters of semi-circular bar, many deformation parameters called sheet resistance, pre-deformation and thickness thinning rate, and so on are measured when the sheet metals are passing through semi-circular bar. And then tensile tests are used to text all sheet metals'yield strength, tensile strength, elongation and subsequent yield strength, subsequent tensile strength, subsequent elongation rate after deformation. A large number of experimental data is accumulated.The way of numerical simulation is adopted to do a large number of simulative analyze that the sheet metals pass the drawbead, and then neural network technology, previous experimental data and numerical simulations are combined to establish predicable model of sheet metal property based on Matlab.On the base of the work of study, BP neural network technology is used to establish a prediction model of subsequent formality of sheet metal after passing the drawbead. This model makes the initial sheet thickness, yield strength, tensile strength, elongation, and the structural parameters of drawbead such as rib height, rib fillet, inner fillet radius and outer fillet radius and so on as the input parameters, and it can predict drawbead resistance, pre-deformation, and subsequent yield strength, subsequent tensile strength, subsequent elongation rate and so on. Compared with the experimental results, it shows that the model has higher prediction accuracy.