Hochpräzise Kunststoffbauteile - verbesserte Auslegung durch druckabhängige Materialdaten in der Prozess- und Schwindungssimulation
Final Report Abstract
The objective to improve the shrinkage prediction of thick walled parts by modification or improvement of material data was accomplished. Although a certain significance for the implementation of pressure dependent data was found, it could be shown that material data in general has to be measured with caution and has to be implemented properly. In the case of thick walled parts especially the influence of the Young’s modulus is strong. The significance of pressure dependent viscosity data could be found for thin parts as well as parts with wall-thickness differences. Because of the lack of reliable experimental data, a new High Pressure Slit Rheometer was designed, which will allow to measure at processing temperatures, with pressures up to 2000 bar and shear rates ranging from the zero-shear -1 viscosity to the beginning of shear thinning (3200 s ). It could be proven that the effect of the measurement procedure (isobaric or isothermal) has a huge impact on the resulting specific volume. However, it was also observed that the difference between the Confined Fluid Method and the Piston Die Method is tremendous. The first procedure varies mainly from the latter in three factors: the measurement of the volume change instead of the absolute volume, the real hydrostatic conditions and the absence of friction during the measurement. Additionally, the data available in databases such as the Material Data Center or from simulation tools is again differing strongly and the measurement method is normally unknown. In the studied case, it was found that the best agreement with the experimental data was obtained from isobaric runs in the Confined Fluid device. The biggest effect on the shrinkage prediction of thick-walled parts is also related to a material property, although it is not pressure dependent. It is the Young’s modulus, because it determines the compliance of the solidified outer layers to the tension pulling them towards the core. Therefore, these values have to be precise especially at temperatures close to the glass transition. As soon as the Young’s modulus was adjusted, the effect of differences in the pvT-data showed an effect. However, with a too rigid surface layer, variations in pvT-behavior were negligible. In this case, it becomes obvious that the models used for the prediction produce good results with good material data. The expansion effect, which was detected in thick-walled plastic parts, is normally not observable in the macroscopic properties such as part dimensions or at least it cannot be separated from opposing shrinkage effects. However, microscopic properties such as the density distribution over part thickness have indicated that this effect exists. Due to the huge volume or rather thickness of the investigated parts in this project such effects are more pronounced because the absolute volume expansion is larger. This effect leads to the deviations between simulation and experiment in the border areas. Assuming the biggest effect of the correct Young’s modulus values in the core as well as the biggest expansion effects at the border, it can be concluded that the simulation would be in good agreement with the experiments if the densification or over-densification could be included. In that case not only the material data has to be measured, but also the model has to be established. Based on these findings, future work should focus on the precise measurement of the change of the Young’s modulus around the glass transition as well as the measurement of the overdensification as a function of formation pressure and cooling time. The latter could be done in existing pvT devices or the development of a new testing apparatus could be approached, which combines the positive effects of the Confined Fluid Method with the ease of use of the non-toxic Piston Die Method. Once the over-densification is known it needs to be incorporated into the shrinkage model by calculating the glass transition within the whole part and coupling it with the formation pressure at this transition. In a next level of difficulty the problem of varying wall thicknesses should again be approached based on the results. However, the observed expansion effect might be negligible in that case, because in most cases (such as convex lenses) the border areas are the thin part.
Publications
-
Effects of Temperature and Pressure on Viscosity during Injection Molding. Proceedings of SPE ANTEC 2011, Boston, MA, p. 1540-1545
Rudolph, N., Pichl, F., Osswald, T.
