Computational Modelling of Metals Extrusion and Forging Processes

"The computational modelling of extrusion and forging processes is now well established. In this work a novel approach is described which utilises finite volume methods on unstructured meshes. This technique can be used to solve simultaneously for fluid flow, heat transfer and non-linear solid mechanics and their interactions. The approach involves the solution of free surface non-Newtonian fluid flow equations in an Eulerian context to track the behaviour of the workpiece and its extrusion/forging, and the solution of the solid mechanics equations in the Lagrangian context to predict the deformation/stress behaviour of the die. Some preliminary examples of this approach will be discussed.IntroductionThe computational modelling of extrusion and forging processes is now well established. Within the last 20 years the finite element method has become the most commonly used technique to simulate metal forming processes [1]-[4]. The Lagrangian and Eulerian type of formulation have been the two major approaches to this problem.Lagrangian methods are efficient and suitable for handling non-linear problems in which small strains prevail, boundary conditions do not change with the course of deformation, and where mesh distortion is not a critical factor in the analysis. Unfortunately, the above points are essential to any accurate metal forming analysis. The problems of mesh distortion and element entanglement pose a serious drawback on the use of such methods. Some automatic remeshing techniques have been developed for Lagrangian finite element analyses of metal forming problems [5]-[7]. However, these methods are not robust and efficient enough to remedy the mesh distortion problems in general. An alternative is to update the mesh manually but this is not always possible, and even when it is feasible it involves major interaction by the user. In addition, remeshing is usually applied to the whole domain which is not necessary and increases CPU time. The complexities associated with remeshing also affect the ability to parallelise the code."