Numerical shape optimization as an approach to extrusion die design

Abstract

Profile extrusion is a manufacturing process used for continuous plastic profiles with a fixed cross section. The key challenge in the development of profile extrusion dies is to design the transition region between outflow and inflow of the die in such a way that the material velocity at the outflow is homogeneous. At the current state of the art, die design is experience based and time-consuming running-in experiments need to be performed for each new die. Aim of this work is to develop a new design approach based on numerical shape optimization with the idea of significantly reducing the number of running-in experiments. Based on an existing, in-house flow solver, a shape optimization framework has been established. It contains a geometry kernel, which operates on non-uniform rational B-splines. The framework has been applied to two profile extrusion dies for profiles with rising complexity: a slit profile and a floor skirting. Apart from validating the functionality of the framework, the aim of the test cases was to investigate the influence of the use of the Carreau model on the optimization outcome. From flow simulations, it can be observed that the use of the Carreau model has a definite influence on the resulting flow solution in extrusion die scenarios. However, it is not clear whether this influence extends to the location of the optimal solution.