Prediction of effective elastic properties of a polypropylene component by an enhanced multiscale simulation of the injection molding process

Abstract In the injection molding process of semi-crystalline polymers, the melt undergoes a complex deformation and cooling history which results in an inhomogeneous distribution of spherulites in the component. To evaluate these inhomogeneities in an isotactic polypropylene (α-iPP) part, an integrated multi-scale simulation approach has been developed in Laschet et al. (2017a) . This approach combines a macroscale mold filling and heat transfer analysis with a crystallization model at the microscale and a two-scale homogenization scheme in order to calculate microstructure dependent elastic properties of the PP component. This approach is extended here by adding some features improving the local Hooke matrix predictions. A link between Molecular Dynamics (MD) simulations and the homogenization scheme is achieved in order to obtain unknown properties of the pure amorphous and crystalline phases. Next, the accuracy of the crystallization model is improved by considering not only isothermal but also athermal nucleation. The spherulite growth is now evaluated by a continuous cellular growth algorithm. Then, several numerical improvements have been integrated in the homogenization tool. The main ones are the implementation of stabilized mixed finite elements to handle accurately the quasi-incompressibility of the amorphous phase and the application of periodic boundary conditions via a master-slave projection algorithm. Eventually, the distribution of effective mechanical properties over the component thickness is compared for different crystallization scenarios with measured elastic moduli.