부산대학교 도서관

The simulation of the injection molding process of semi-crystalline thermoplastics is widely used and commercially available. Typical use cases for process simulations are melt flow optimization, tool design and finding process parameter sets for the production. In recent years, injection molding simulations also provide the prediction of shrinkage and warpage. These simulations, however, neglect the variations in local mechanical properties caused by inhomogeneous solidification and the resulting locally different spherulites microstructure. In order to account for the locally inhomogeneous solidification and the resulting local microstructure, and hence to improve the precision of the warpage prediction, a multi-scale simulation chain is developed and presented here. The multi-scale simulation chain predicts the shrinkage on the part scale taking into account the mechanical property distribution, which is derived based on the local microstructure using a two-stage numerical homogenization scheme. The numerical homogenization is able to consider the local degree of crystallization, which is noticeable reduced in the surface layers of the component, resulting in a reduction of the mechanical properties. To show the applicability of the multi-scale simulation chain, an experimental validation is presented using injection molded plates with thickness steps. The final warpage is measured using a tactile measurement (Zeiss O-Inspect 442). The bending moment caused by warpage and relative magnitude are compared with the numerical results. [ABSTRACT FROM AUTHOR]