Simulation of Dynamic Effects in Progressive Die Operation and Control

Abstract

The demand from the automotive industry for increasingly complex sheet metal components and higher throughput in progressive die operations has led to the increased integration of sensors and control-systems in the sheet metal forming process. However, current control-systems used in sheet metal forming are often limited to measuring the state of the tooling during the forming process, neglecting the dynamic effects of the strip during its transfer between tooling operations. Developing a control strategy that accounts for the strip dynamics requires knowledge of how various process parameters influence the strip behaviour during both the transfer and forming stages. FE element models can accurately model the behaviour of sheet metal, but by themselves cannot identify a robust control strategy. Machine learning can solve this issue by constructing a probabilistic representation for the data generated from FE simulations to be used to identify a control strategy for sheet metal forming. The goal of this work is to conduct a parametric study on a progressive die FE model and evaluate the influence of various input parameters. The data collected from this FEM study will be used to construct a neural network model that will inform a control strategy for a progressive die.