The optimization of an extrusion process is a conflicting, multi-objective problem. It is complicated by the number of variables (screw/die geometry, operating conditions, material data) and their non-linear relations, as well as by the opposing criteria, for example extrusion throughput and power consumption. It is difficult to find the global optimum for the process avoiding local optima. There are two approaches to solve the problem, experimental and using a mathematical model of extrusion. Optimization techniques based on an experimentation are time-consuming and very expensive. In this paper we present an optimization methodology based on the Genetic Algorithms (AG), where response surface is given by the extrusion model. A mathematical Single-Screw Extrusion Model SSEM developed at the Warsaw University of Technology is used to predict the extruder behavior, and AG approach is used for optimization. An integrated SSEM-AG system was developed to study optimization of the single-screw extrusion process. Three design criteria (output variables) are selected for optimization: maximum extrusion throughput, minimum power consumption and low melt temperature. As input variables, screw speed, barrel temperature and screw channel depth are chosen.
Residence time distribution of a pharmaceutical grade polymer melt in a single screw extrusion process