Springback prediction in sheet metal forming: constitutive equations, finite element simulations and experimental validation

Predictive methods appear to be the most effective way to solve springback in sheet metal forming. The accuracy of the predictions depends upon the application of accurate material modelling. Experimental devices and methods are being continuously improved to incorporate increasingly accurate plastic bending characteristics. As part of these efforts, a new tool has been developed to test and record the characteristics of sheet metal deformation by investigating the Bauschinger effect factors (BEF) and the identified hardening parameters. The developed tool is believed to simulate the actual forming conditions of bending and provide more reliable information. The initial experimental investigation shows that the Bauschinger effect does occur during bending and unbending loadings in sheet metal forming. The BEF value was found to increase as the thickness increases. Therefore this justifies the need to consider the Bauschinger effect in sheet metal forming simulation through the use of relevant constitutive equations. A direct optimization method has been successfully applied to identify material hardening parameters from the acquired experimental data of the newly developed tool. The optimisation result shows that nonlinear kinematic hardening and nonlinear mixed hardening models are capable of fitting the smooth transition curve of the experimental hardening data. Mixed hardening model performance however is considered to be much better as proven by lower residual or fitting error values. This justifies the idea that the application of a mixed hardening model is more suitable for springback simulation in sheet metal forming. Validation work was conducted in order to test the effectiveness of applying the two hardening models by incorporating the identified parameters in predicting springback using finite element simulation. Of the two, the mixed hardening modelling has been proven to provide better simulation results in predicting springback