Open die forging process simulation: a simplified industrial approach based on artificial neural network

Abstract

Simulations by Finite element analysis (FEM) of open die forging process related to different configurations are quite common in industry to optimize the process. This approach, anyway, is relatively slow to be performed: hence it is not suitable for online optimization of the forging processes. In this paper a simplified approach is proposed aimed to describe the plastic strain at the core of the forged component. The proposed approach takes into account the plastic deformation at the core of the forged component and consists on a thermo-mechanical FEM model implementation allowing to define a set of equations giving as output the plastic strain at the core of the piece as a function of the forging parameters. An Artificial Neural Network (ANN) is trained and tested aimed to relate the equation coefficients with the forging to obtain the behavior of plastic strain at the core of the piece.