Effect of different contact formulations used in commercial FEM software packages on the results of hot forging simulations

  • T. Hatzenbichler
  • O. Harrer
  • B. Buchmayr
  • F. Planitzer


Commercial FEM-software packages are widely used in the industry to predict material flow, temperature
distribution and die load during the forging process. Contact in conjunction with plastic material behaviour,
which is typical for forging simulations, leads to highly nonlinear equations in the FEM algorithms, which
may cause problems in numerical convergence. Some FEM software providers handle this problem by
automatic contact damping or similar algorithms. However, the user has mostly no detailed information about
adjustments and prediction accuracy. The only possibility for the user to have an impact on the contact
behaviour is to set a friction factor and to choose a friction model (e.g. Coloumb or Shear) appropriate to the
investigated process. Friction factors are often measured by standard tests like the ring compression test which
should be valid for all used software packages. In this paper a benchmark between three software programs is
performed based on a model for ring compression tests under typical hot forging conditions. The commercial
FEM-software programs Deform2D, Forge2007 and Abaqus are compared by generating a nomogram for each
software package. For all simulations identical physical (temperature, flow curves etc.) as well as numerical
influence parameters are used. The simulations show a significant divergence in the results depending on the
used FEM-software. This leads to the conclusion that a friction coefficient which is true for one software
package can not be transferred directly into another one.