2D and 3D simulation of complex multistage forging processes by use of adaptive friction coefficient

Abstract Tool stresses, especially in multistage cold forming processes, are mostly very high. They are limited by the maximum permissible stress of the tool. For a correct tool design it is necessary to analyse the stresses at the tool–workpiece interface with a high degree of accuracy in order to use the process at the limit and to avoid an early tool damage. In this analysis friction plays a central role because it is strongly influenced by the existing distribution of the contact variables at the tool–workpiece interface. Moreover these contact conditions have a great influence on the material flow. For its realistic prediction, e.g. by using the finite element method (FEM), it is necessary to obtain a realistic relation between these varying local contact conditions (normal contact pressure, sliding velocity and temperature) and the friction coefficients. For this purpose we developed a new method to include local friction parameters in FEM calculations. The basic values will be obtained by a ring compression test, which is performed under similar conditions (surface roughness, lubrication, stress configuration) as the investigated process. The friction coefficient will be evaluated by a FE simulation of the ring compression test using measured displacements as boundary conditions. The dependency of the friction values on the contact parameters will be calculated by neural network techniques. Transferring these values to a FE simulation of an investigated process and taking into consideration their dependency on the contact parameters we achieve a higher degree of simulation accuracy. A further increase of simulation accuracy will be obtained by the use of nearly undeformed FE elements at the contact area. This demands special remeshing techniques and hexahedral elements in the case of 3D simulations.