Prediction of wall thickness in deep drawing process with neural network

In this paper, the modeling of deep-drawing process using neural networks is established. The relationships between process parameters (punch radius, matrix radius, blank holder force) and part quality (wall thickness) are created, based on a neural network. Finite element analyses are conducted for combination of process parameters designed using statistical full factorial experimental design. A predictive model for wall thickness is created using Levenberg-Marquardt (LM) artificial neural network exploiting finite element analysis results. The results obtained are found to correlate well with experimental data.

Keywords:

sheet metal forming , full factorial , Finite element method , Artificial neural network