Industrial Fe-Simulation of Roughing Using an Automatic Solver Shifting Technique

Abstract

Roughing has been simulated with the Finite element software AbaqusTM to replicate an industrial-scale process. The model has been made to be as close as possible to its real counterpart. For this purpose, an automated controlling logic has been created to simulate the multiple passes as well as inter-pass times for roughing. Simulating multiple passes with FEM is computationally very demanding, so new methods to reduce computing times are worth considering. During a roll pass an explicit solver is necessary due to high deformation amounts and rates. An explicit solver is tied to a very small time increment, so it takes a long time. On the other hand, inter-pass periods do not include any deformation or roller contact, so an implicit solver is quite capable of computing this portion of the simulation. An implicit solver can speed up the time increment considerably when compared to the explicit solver, so using it potentially saves a significant amount of computing time. Unfortunately, Abaqus does not include any methods to change the solver during a single simulation. Instead it is possible to communicate between the two solver types by manually importing data from a completed simulation to a new simulation model. A new method to change solvers automatically using a self-made Python code is proposed in this paper.