| Hi-index | 0.00 |
In thin sheet metalforming process, the main cause of dimensional errors is the springback. The errors level is very responsive to small variations of the blank features (such as thickness or mechanical properties) and of the process features (such as the friction between blank and the blank holder). The measurement procedure to be applied in order to check if thin sheet workpieces are complying with the dimensional accuracy requirements implies higher costs than the procedure to be applied to thick sheet workpieces. The present paper proposes a springback optimal assessment technique based on support vector machine classification in which the training data are obtained by Finite Element Method (FEM) simulations. The trained model is used with data obtained from a real thin sheet metalforming process for optimal dimensional assessment. The optimization criterion is the assessment performance and the restrictions considered are the online assessment, along with small number of measurements and fast assessment procedure. The simulation of proposed springback assessment technique confirmed its efficiency. Furthermore, in some conditions, the assessment may lead to predictive control.