Band 12, Heft 3

The Gas-Assisted Injection Moulding process (GAIM) offers a number of advantages, which are almost without exception related to the negligible pressure drop in the gas core of the moulded product. Accurate prediction of the gas distribution inside GAIM products is, however, still a major problem to be solved. As gas penetration is governed by three-dimensional phenomena, one has to resort to numerical simulations to analyse the process. Therefore, a numerical model for GAIM simulations has been developed, based on a physical, rather than on an empirical approach. The model employs a pseudo-concentration method in order to avoid elaborate three-dimensional remeshing, and has been implemented in a finite element program. Simulation results for two-dimensional test cases representing typical GAIM situations prove that the model covers most of the important aspects of gas injection. A pilot simulation for a simple three-dimensional mould demonstrates that the model is able to deal with three-dimensional GAIM. Further refinement of the model will mainly concern the implementation of faster and less memory-consuming solution methods.

The versatility of the injection moulding process can be increased by combining multiple polymers in one product. The multi-component injection moulding process (MCIM) offers the possibility to inject two or three materials sequentially or simultaneously into a mould to make products that consist of e.g. a layered structure. For a successful application of this technique for geometrically complex products or multiple materials, the material distribution in the product has to be predicted. In this paper numerical tools are described for calculating the positions of material particles during the flow in the mould cavity. These tools make it possible to solve the inverse problem of predicting the injection sequence in MCIM given a required material distribution in the product. The simulations are validated using experimental results of a co-injected strip with stiffener ribs. Furthermore the effect of a bifurcation of the midsurface on the material distribution is investigated numerically using a simplified model based on the local mass balance.

The prediction of shrinkage and warpage of crystalline polymers is quite difficult. This is because of the complications of the crystallization process and associated material property changes. In typical injection molding of semi-crystalline polymers, it is difficult to calculate crystallinity accurately because of its complicated dependence on temperature, time and stress. In the case of slowly-crystallizing polymers, typically the crystallinity of the polymer remains small during the process. In this case, therefore, the importance of accurate calculation of crystallinity is decreased. The difficulty in this case is the measurement of material properties under variable crystallinity because of the possible crystallization during measurement. In the present study, methods have been developed to obtain material properties which include the effect of crystallinity. The material properties measured include viscosity, thermal conductivity, heat capacity, PvT relation and crystallization kinetics. Injection-molding experiments have been conducted to measure the shrinkage of a part using a slowly-crystallizing polymer. Shrinkage has been measured by comparing the mold cavity dimension and the dimension of the molded sample. The polymer used in this study is PET (polyethylene terephthalate). A simulation program has been developed to analyze the injection-molding process. This program is based on quiescent crystallization kinetics. The predicted shrinkage and measured shrinkage are found to agree reasonably well.

Fiber filled injection-molded plastic parts contain complex fiber orientation patterns. This fiber orientation state affects material properties, including both elastic modulus and strength, and the calculation of shrinkage and warpage. Several models are available for predicting fiber orientation, which are based on the theory of Folgar and Tucker [1]. A particular simplification of these models, here called the quasi-planar (QP) approximation, was introduced by Gupta and Wang [2]. Here we develop a substantially improved version of that model, which we call the optimized quasi-planar (OQP) approximation. The OQP approximation is formed by optimizing the QP approximation against a more general fiber orientation model. Details of the current and improved models are discussed, and the typical behavior of the old and new models is explored. In addition, simulations are performed using C-MOLD, a commercial software package that implements the QP approximation, to determine the final fiber orientation state in an end-gated strip. The resulting fiber orientation data is then used to predict the elastic modulus as a function of position, and both predictions are compared to available experimental data for fiber orientation and elastic modulus. The results show that the OQP approximation improves the fiber orientation prediction and the subsequent elastic modulus prediction compared to the QP model, though the OQP approximation is still less accurate than the full three-dimensional model.

Making use of available data from the literature, a master correlation based upon a two-domain-Tait representation is developed which describes the asymptotic high-temperature and low-temperature PVT behavior of isotactic polypropylene. The transition between these two asymptotes in the cooling mode is then handled based upon the quiescent-crystallization-kinetics equation of Nakamura, employing model constants for the temperature- and pressure-dependent crystallization rate based upon earlier work by the present investigator. As a further step, it is shown that a rule-of-mixtures approach can describe the low-temperature and high-temperature asymptotic PVT behavior equally as well as the 2-domain-Tait model while requiring two less free parameters. In turn, since the rule-of-mixtures representation contains the absolute crystallinity as an explicit parameter, it is proposed that this model might also be used under flow-induced conditions once the corresponding crystallization kinetics become established.

Commercial injection mold filling simulations use the lubrication approximation to model pressure and temperature during flow. This approximation is valid when flow streamlines are parallel to a cavity surface, such as when the cavity thickness is much smaller than the flow length. Yet this approximation may not be valid in complex cavities where the streamlines are not parallel. Flow through junctions, gates and sprues are always accompanied by a velocity component transverse to the mold surface which is not included in the model. Neglecting transverse flow through use of the lubrication approximation can result in anomalous changes in bulk mean temperature, filling patterns, and molecular and fiber orientation. This paper presents a mass balance algorithm for approximating transverse flow through gates, sprues and junctions. This new algorithm is presently suitable for only relatively simple geometries. Future work will attempt to make a similar algorithm for transverse flow through complex mold shapes with negligible penalty in additional computational run time.

This paper presents an effective numerical model capable of simulating the filling stage of the gas-assisted injection molding process, particularly the gas penetration phenomenon involving the gas-polymer interaction. A Galerkin Finite Element model is used to model the polymer flow during the filling and gas penetration stages. The polymer flow front and the gas-polymer interface are tracked using a volume tracking technique. Finally, a finite difference technique has been used to calculate the temperature across the thickness. Comparison between numerical simulation predictions and the experimental data for a plate mould as well as for a complex part is presented.

A non-Newtonian flow model for individual elements of a modular tangential counter-rotating twin screw extruder is developed. A non-Newtonian non-isothermal flow model for a composite modular machine is then described. Fill factor, pressure and temperature profiles along the axis of the twin screw machines were predicted for various modular screw configurations. Finally, experiments were carried out on the modular machine to verify the predictions. Generally, good agreement with the flow analysis was found.

A numerical procedure to simulate the formation of a sink mark near the base of a rib in an injection-molded plastic part has been developed. This method uses two commercially available software packages, C-Mold and Abaqus, to model the injection molding process. The C-Mold software uses a generalized Hele-Shaw formulation to simulate the mold filling, packing and cooling stages of injection molding. From this analysis, pressure and temperature histories in the rib area are derived. These are used to determine the initial conditions and boundary conditions for a sequentially-coupled thermal / structural finite element analysis of a cross-section of the rib, using the Abaqus software. An elastic-plastic material model with temperature-dependent material properties is employed. The thermal/structural finite element analysis provides an accurate model of the two-dimensional cooling and shrinkage behavior of the part during molding. To verify the accuracy of the method, the dependence of sink mark depth on various geometric and molding parameters has been analyzed. The predicted sink mark depth is found to be in good agreement with experimental results from the literature.