Band 8, Heft 4

In most polymer processes the quality of the final part stems back in great part to the mixing of the polymer blends and the viscous heating that occurs during the process. To date, these processes have been simulated using the finite difference or finite element methods which rely on cumbersome remeshing techniques, restricting the analysis to simple geometries. However, mathematical manipulation of the governing equations results in boundary integrals, allowing the solution of complicated moving boundary problems without domain meshing. The resulting boundary element model and simulation results of flow with heat transfer and the blending of two or more fluids of different viscosity effects during mixing of polymer blends is presented in this paper. The results give insight to the complex phenomena that take place during mixing processes.

The elementary processes during heat transfer in screw extruders are described by a simulation model. The balance equations are solved numerically with a personal computer. By contrast with known computational approaches, consideration is given to the non-stationary processes in the melt layer between the screw flight and the barrel wall. This is guaranteed by incorporating the “pulsating shear” in this layer into the model as a dissipative energy source. The computational results are used to demonstrate, in an exemplary fashion for one extruder series, the influence of all the relevant machine and process parameters on the internal heat transfer.

Reactive extrusion is the term used to denote a plastics processing method in which an extruder is used as a reactor for the continuous synthesis and modification of polymers. This paper describes the manufacture of polystyrene and styreneisoprene copolymers by a reactive extrusion process: anionic “living” s-butyl lithium initiated bulk polymerization was performed in a co-rotating closely intermeshing twin screw extruder. The results of the process analysis show that living polymerization of styrene can be performed in a screw-type reactor, despite the high reaction temperatures of over 200 °C. The polystyrene melt can be modified in bulk with comonomers or coupling reagents immediately after polymer synthesis. Depending on the raw material, formulation, and process parameters, the process variants developed and analyzed (homopolymerization of styrene, copolymerization of styrene-isoprene mixtures, sequential polymerization of styrene and isoprene) resulted in styrene polymers with widely differing structural characteristics and properties. For example, the copolymerization or styrene-isoprene monomer mixtures produced poly [(isoprene-co-styrene)-b-styrene]. The sequential polymerization of styrene and isoprene led to poly(styrene-b-isoprene) contaminated with partly crosslinked low molecular weight polyisoprene. The polyisoprene content is presumably formed by side reactions due to the high reaction temperatures.

A long-fiber injection molded phenoxy composite has been developed by using a new processing technique that combines reaction injection molding (RIM) and pultrusion to produce a prepreg pellet. RIM-pultrusion uses a reaction injection system instead of the resin bath typically used in pultrusion. The resin system and process developments are described. The fiber damage during injection molding is minimized by using the RIM-pultruded pellet as prepreg. The resulting long-fiber carbon fiber composite has an average volume aspect ratio that is an order of magnitude higher compared to typical short-fiber injection molding. The fiber length distribution (FLD) is measured for both glass and carbon composites. The extent of bundling has been determined from the moment ratio. The extent of reaction of the diepoxide during processing is measured using Fourier transform infrared spectroscopy (FT-IR). The pultrusion line speed and temperature are determined for secondary processing, namely injection molding, from the epoxide conversion. Also, dynamic mechanical analysis has been performed on the resin, prepreg, and injection molded sample to study the mechanical and thermal properties. Scanning electron microscopy is employed to examine fiber-matrix adhesion, void content, bundling, and fiber pull-out.

The flow characteristics of Newtonian and non-Newtonian fluids in a roller die are simulated. Both pre-form and Tmanifold entry regions are considered. Calculations are made using a modified Flow Analysis Network procedure. We take into consideration the interaction of roller dies with the screw extruders that feed them. We show operating conditions that can cause starvation to develop in screw extruders feeding roller dies.

The simulation method is able to relieve the engineer of experimental work during the planning of extrusion plant and thus clearly reduces the development time. Sufficiently precise determination of the calculable parameters is possible with the process models developed so far. The empirical knowledge of the engineer is, of course, necessary as well in order to estimate the impact of influencing parameters that are not taken into account in the calculation models. The presented simulation program REX can perform these complex calculations in a comprehensive manner, with its simple and comfortable user interface. The program, however, can be employed as a valuable tool only if it is in the hands of an expert operator who can, and indeed must, be able to interpret the results as well.

This paper presents an application example for the modelling of a mold used for the gas-assisted injection molding process. The simulation software used allows the computation of the distribution of the thickness-fraction of skin material all over the part depending on the process conditions. The illustrations are shown in connection with an industrial mold and the simulation results are compared with experimental data showing a reasonable agreement. The influence of variations of different operation point parameters is shown. Some effects occurring due to the gas flow into the cavity are described.

This paper presents the simulation of a typical cyclic polymer processing operation with respect to the performance of different control strategies for part quality when subjected to disturbances with realistic load-profiles. Disturbances enter the simulation in a complex, probabilistic manner creating a unique flexibility in load representation which helps the simulation more accurately capture the performance of real processing operations. Realistic simulations can provide necessary information and be less costly than benchmark studies on actual processes. The objective of the research is to examine the effect of different control strategies on the overall control of part-to-part quality, not the control of the continuous processes that occur in part fabrication. The performance of six different control strategies is compared; two strategies are conventional feedback, and four strategies are statistically based. The statistically based algorithms effect closed loop control, but only when a true load exists. Results are analyzed using statistical analysis of variance (ANOVA) via a completely randomized block experimental design, and Duncan's multiple range test is used to rank the control strategies. The results indicate that as more disturbances enter the system, the conventional controllers perform better than the other strategies. When fewer disturbances are present, however, the statistically based controllers perform better. The most notable result is that one statistically based controller, the Western Electric runs rules controller, performs well over the entire range of disturbances.

A new method of flow visualization has been used to film the injection mold filling operation. A study of jetting has been performed using real-time neutron radiography which permits real-time visualization through metal molds. Results of this method for a thermoplastic elastomer filling an end-gated strip mold are reported for a bench-scale injection molder at constant pressure.

The paper describes a series of blow molding trials on two different polymers under a variety of processing conditions. The tests were performed to investigate factors that affect the variation of wall thickness in a blow molded part. In addition, a numerical analysis was performed to predict the wall thicknesses. A major outcome of the tests was that the relative thickness distributions in the parts were independent of processing conditions and of the material used to form the part. The primary influence of the processing conditions was to put more or less material into the part, but they did not significantly affect the relative distribution of the polymer. The analysis gave a good overall prediction of wall thickness distributions.