Analysis and optimization of mixing inside twin-screw extruders

The most commonly used mixing device in polymer processing is the closely intermeshing, co-rotating twin screw extruder. The main goal of this project is mixing simulation inside twin screw extruders using mapping method. This project consists of two parts; three dimensional velocity field simulation using finite element method combination of extended finite element method , and mixing analysis using particle tracking and optimization with mapping method for different configurations of screw elements. Accuracy of mixing analysis depends on accuracy of velocity field. Determining the velocity field in twin-screw extruders operating in realistic conditions is still very challenging despite all progress made in simulations during last years. There are two main difficulties that arise when solving the balance equations. The first is that we have to deal with moving geometries, the screws, in a fixed barrel and the gap widths are extremely small and change in time during rotation. Obtaining a velocity field and its derivative very close to the moving screws is not trivial, but essential for further particle tracking analysis. A second challenge is to deal with the viscosity that is a function of position, shear rate and temperature, and can change orders of magnitude. In this work we present an extended finite element method (XFEM) to model non-Newtonian Stokes flow inside the twin-screw extruder and demonstrate its accuracy and efficiency by systematically refining the mesh and compare with boundary-fitted results. Two-dimensional cross-sectional results for velocity and pressure are compared with full three-dimensional simulations. Residence time distributions are compared for various screw designs. Analysis and even optimization of mixing based on particle tracking is far from trivial since no volumetric data is available. In this work we continue on the application of the mapping method, which does provide volumetric quantities, to quantitatively compare different screw layouts and find optimal designs. The mapping method has proven its merits for static mixers where, in previous work, all popular static mixers are evaluated and guidelines have resulted to design new mixers based on either compactness of minimum pressure drop. In this work we have used to same approach to analyze mixing and we have compared several screw configurations and different types of screws. In particular, conveying elements are compared with different pitch length and gap width, while for kneading elements a variety of staggering angles have been put side by side. The flux-weighted intensity of segregation is used a mixing measure to evaluate the different screw designs.