Multiscale Constitutive Modeling of the Mechanical Properties of Polypropylene Fibers from Molecular Simulation Data

We present a multiscale approach to create a constitutive model that predicts the mechanical properties of polypropylene fibers based on chemical and physical characteristics. The development of this method relies on validation with experimental stress-strain curves from nine different isotactic polypropylene (iPP) fibers with their varying molecular weight characteristics, Hermans orientation factors, and crystallinity. Complementary molecular models were built by using molecular dynamics (MD) simulations with united atom models. Tensile deformation simulations adapting a quasi-static procedure resulted in stress-strain curves that aligned well with the experimentally measured ones. A neural network model was trained on the MD simulation data to create correlations that predict parameters for a chosen constitutive model that describes the mechanical properties of the polypropylene fibers. This computational approach is amenable to be applied to polymer fiber systems and aims to aid in the design of polymeric materials to achieve targeted mechanical properties.

Automated summary

In this paper, a multiscale approach is presented to predict the mechanical properties of polypropylene fibers based on chemical and physical characteristics. The method relies on validation with experimental data and utilizes molecular simulations and neural network models to predict parameters for a chosen constitutive model.