Design Optimization for Cold Forging by an Integrated Methodology of CAD/FEM/ANN

Abstract

In this paper, an integrated methodology of geometric modeling, finite element method (FEM) and artificial neural network (ANN) is proposed for design optimization of cold forging process. Forging processes with some key design variables are firstly simulated using rigid-plastic FEM so as to create training data for the ANN model. Then neural network model is used to predict for unseen data after being properly trained. Finally, some forging experiments are carried out to confirm the ANN prediction results, a good agreement is found between the predicted data and the measured results.