Volume 15, Issue 2

Mixing is a major ingredient in many industrial processes to obtain desired and uniform material properties. Nowadays, many materials, pigments, additives, gas or reactants, are mixed to create new products combining the characteristics of different raw materials to obtain specific product properties. The quality of the mixing, i.e. the uniformity of the mixture, is a key issue that will determine the morphology and the properties of the resulting compound [1, 2]. An insight in the physics of mixing is therefore necessary in order to achieve a good quality of mixing or maintain it when scaling equipment, for example. Such information can now be obtained through the numerical simulation of the transient flow in extruder components. To improve greatly the ease of obtaining such information, a new technique introduced in a finite elements software is presented. This technique simplifies the meshing, reduces the meshes needed and eliminates complex remeshing algorithms to simulate flow in screws, pumps and mixing devices. This technique is validated versus traditional simulation methods (i.e. conforming meshes). 3-D transient numerical simulations of two twin screw extruder configurations are then presented. A further quantitative comparison of their mixing behavior is developed as statistical information (of the RTD, deformations, dispersion, etc.) can be obtained to compare both configurations in a synthetic and quantitative way.

Computational fluid dynamics (CFD) based on a finite-volume method was used to simulate three-dimensional pressure and velocity fields for non-Newtonian fluids in fully filled screw elements in a co-rotating twin screw extruder. The employed method also allows the calculation of temperature fields for non-isothermal flow which requires a non-steady state calculation. The CFD code was first evaluated on simple flow cases such as flow of a Newtonian or non-Newtonian fluid through a cylindrical die or flow in single screw extruders equipped with different screws. The results agreed very well with analytical calculations (die flow) and experimental results (single screw extruder). In a next step, we wanted to establish the reliability of the numerical results for twin screw extruders by performing experiments on a 40 mm and a 58 mm screw diameter machine. Temperature and pressure distribution along the screw in the fully filled section in front of the die were measured. Melt conveying in several screw elements (different pitches) was studied for different materials (amorphous polystyrene and partially crystalline polypropylene) at various flow rates and screw speeds. It was attempted to measure axial pressure build-up for flow rates from nearly zero to beyond the drag capacity of the specific element. The measured bulk values of pressure were then compared to the corresponding values extracted from the CFD simulation results. So far, only isothermal calculations (at different temperatures) have been performed. The temperature profile along the screw was determined by calculating the energy dissipation in isothermal slices and converting the dissipated energy into internal energy (temperature increase, neglecting heat losses at the wall). General agreement between experimental and numerical results is good.

Grooved feed extruders have been around since about the 1960s. These extruders offer considerable advantages over conventional extruders, such as higher throughput, better stability, and the ability to process very high molecular weight polymers. There are some important disadvantages as well, for instance, higher motor load, wear is more likely, high pressures in the grooved region, and the screw design has to be adapted. The disadvantages of the grooved feed extruder disappear when the grooved feed extruder is made with a mechanism that allows adjustment of the groove depth. This paper will report on the development of grooved feed extruders that incorporate an adjustment mechanism that allows the depth of the grooves to be changed, during actual operation, from zero to full depth. Operational data from actual extrusion experiments are presented.

Continuous hydrolytic depolymerization of polyethylene terephthalate (PET) resin at high temperature and pressure was carried out in a co-rotating twin screw extruder. In this novel process, hydrolysis is achieved by injection of saturated steam at high pressure. Optimum conditions occur at relatively low screw speeds and water/PET ratios. Hydrolysis extent was measured using carboxyl end group titration, and by determination of molecular weights using NMR and GPC analysis. Low molecular weight products are obtained at low residence times in the extruder, suggesting high depolymerization rates. This is supported by the results of a simple kinetic model.

The use of a drag flow device such as a twin screw extruder for polymerization and/or copolymerization processes has been studied extensively over the years. Due to the high viscosities (100 to 100 000 poise) that are produced, poor backmixing is experienced in these drag flow devices. As a result, these studies have shown the extruder reactor to exhibit a ‘‘plug flow’’ behavior. A drag flow device has been developed that achieves good backmixing with high viscosity materials. This device incorporates two counter-rotating, nonintermeshing, screws; one screw conveying material forward, while the other conveying material backward. Theory suggests that in a plug flow reactor the copolymer composition varies (drifts) as a function of time (position) in the reactor in most cases. Conversely in a micromixed continuous stirred tank reactor (CSTR) the copolymer produced does not drift as a function with position in the reactor. It has also been shown that this drift in copolymer composition becomes more pronounced at higher conversion. Experiments were performed in a counter-rotating nonintermeshing twin screw extruder using the ‘‘backmixed’’ configuration and a ‘‘conventional’’ configuration. Two copolymer systems were investigated; butyl acrylate/styrene and butyl acrylate/butyl methacrylate. The copolymerizations were conducted at three residence times for each system and configuration to provide materials of a wide variety of conversions (13 to 97 %). Monomer samples were taken at two ports in the extruder and at the die. Samples were analyzed for composition by two methods; FTIR and refractive index (RI). It was shown that the copolymer composition of samples produced in the backmixed extruder reactor does not change significantly as a function of position in the reactor. This is an indication that the copolymer composition does not drift in the reactor. Theoretically this is the behavior that exists with a chain addition copolymerization that occurs in a CSTR. Two methods (RI, FTIR) were used in the copolymer formulation (BA/Sty) and one method (FTIR) was used in the other copolymer formulation (BA/BMA). It was further shown that the difference in copolymer compositions in the two extruder reactors is more pronounced at higher conversions than at lower conversions.

An experimental study of film formation and structure development of polybutylene terephthalate (PBT) in single and double bubble tubular film extrusion processes is presented. PBT was found largely stable in the first stage. In the second stage, the bubble was unstable and biaxial inflation was achieved only with a blow-up ratio of one. The sources of second bubble instabilities were related to the crystalline nature of the first bubble being inflated in the second stage. The structure and properties in the films varied substantially with processing conditions. While the first bubble had poorly ordered α-structure, the double bubble possessed oriented polymorphic texture. The second deformation that induced high film-line stress favored the formation of a stable β phase. The mechanical behavior of the films reflected the crystal perfection as well as orientation in the films.

In this paper, the viscoplastic material model of PET (polyethylene terephthalate), which is intended to be used in the FEM (finite element method) simulation of stretch blow molding process, has been studied. Material tests of PET were performed with the constant strain rates varying from 0.01 to 1 (1/s), at temperatures ranging from 90 to 150 °C, based on the obtained data a two-stage model was proposed. The proposed model could precisely take into account the effects of strain hardening, strain rate sensitivity, variation of the hardening index, and temperature dependency. This model has been implemented into the nonlinear finite element code PBLOW3D, which is developed in the Riken, and its performance in the stretch blow molding simulation has been studied. It has been demonstrated that the proposed material model provides significant improvements, compared with two existing material models, in the simulation of the blow molding process of PET bottles.

Injection compression molding incorporates a mold compression action in the postfilling stage to compact the materials. The process is the same as the conventional injection molding in the filling and first stage packing. Packing stage continues until the gate material is frozen or the gate is shut off. Then, a mold compression action is engaged and continues to the end of the molding. With the mold compression, the pressure distribution inside the mold cavity can be uniform and the residual stresses caused by pressure gradients can be minimized. However, due to the non-isothermal cooling process, the compressibility of each place inside the mold cavity was not the same. At some time of the cooling process, the pressure at different points began to deviate from each other. In order to improve the non-uniform compression at later the stage of a ICM process, a two-stage compression method can be used. A disk mold was designed to study the effects of the two-stage compression on the molding product. It was found that less warpage can be obtained with a lower compression force at the second stage. The experimental results also suggested that there existed an optimal compression time that was corresponding to the time before the material solidification.

In this paper we describe a model to predict crystallinity from thermal history while compression molding UHMWPE. A thermal model is used for temperature predictions within a compression molded disc. The temperature predictions are used in a residual stress model (see Part II of this paper). Three different process cycles were studied, one with fast cooling, one with slow cooling and one with soak at crystallization.

In part I of this paper a thermal model was used for temperature predictions within a compression molded disc. A process simulated laminate technique was used to measure residual stresses. The fast cooling cycle produced high residual stresses (∼10 MPa). The slow cooling cycle produced low residual stresses (∼1 MPa). The cycle with soak at crystallization also produced low residual stresses (∼2.6 MPa). The modeled and measured values were in close agreement. A major portion of the residual stresses (∼70 %) was developed within a short time period (∼10 min) after the end of crystallization for all the cycles.

The mechanical properties of injection moulded unreinforced isotactic, polypropylene depend on the crystalline structure and lamellae orientation. Push-pull processing is an injection moulding technique that allows the user to produce parts with oriented structures. Mostly applied for short fibre reinforced thermoplastics in order to align fibres parallel to the flow direction, this technique can also be used to orient the polymeric matrix of LCP and polypropylene. A comparison of mechanical properties as well as investigations on morphology between the samples produced using conventional injection moulding and push-pull processing are presented trying to improve the knowledge about this so far rather unknown technique.

This paper lays out the main procedures for performing morphological characterisations of Long Glass Fibre (LGF) composites with particular reference to the sheet extrusion and thermoforming processes as they may be configured for production. The techniques used, including optical microscopy, scanning electron microscopy and image analysis, are described both with regard to their laboratory application to these materials and to their potential for monitoring the performance of the industrial manufacturing process. Results obtained from the different techniques at the three various stages of the manufacturing route are presented and discussed in terms of the structure property relationships obtained and the reinforcing fibre mat system typical of these types of material.