Modelling stress-strain curves and forming limit curves using deep learning

Abstract The sheet metal forming process stands one of the most principal applications in production engineering with the goal of producing defect free products possessing the mechanical properties that met the working conditions requirements. The efficient metal forming process relies on a deep understanding of the material behaviour during the forming process. To achieve this, feasibility studies for new incoming components helps determine if a part is feasible for manufacturing or if design modifications are required. This simulation requires an accurate material card to replicate the real material behaviour, requiring extensive material testing data such as stress-strain curves and forming limit curves (FLC). Artificial intelligence, specifically deep learning, is a promising tool to overcome such challenges. In this study, one of the deep learning algorithms, the artificial neural network (ANN), is utilized to create material testing data for building the material cards for feasibility simulation.