Advances in digitization allow the rapid transfer and processing of large data sets and will thus enable digital twins of complex systems that can be used for analysis. New methods for linking measurement data and physics-based model predictions are developed using the example of lifetime prediction of an eBike with SFRP frame [1,2]. In this project, measured data on thermomechanical and electrical loads is collected on a large scale, stored in the data cloud and used for life prediction using methods of artificial intelligence.
In order to feed these data-driven surrogate models, broad experimental data as well as physics-based models for predicting static and fatigue behavior of the SFRP frame structure are required and offered here. In this contribution the characterization of the SFRP material (IXEF© 1022, Solvay Specialty Polymers) under quasi-static and fatigue loading is presented. Specimens with varying thickness and fiber orientation are manufactured on coupon level and tested under different temperatures. A special focus is put on the relationship of microstructure and properties. For this, the local fiber orientation tensors are calculated from micrographs in terms of image correlation with an improved practice-oriented methodology based on [3].
Mean field homogenization by means of the Double Inclusion Model [4] is used to describe the nonlinear material behavior for arbitrary fiber orientations and is validated against the experimental data on coupon-level. For modelling the process-structure-property relations on structural level a coupled FE-simulation is suggested which links the derived material model with the process simulation of the molding process of the thermoplastic bike frame. In this context static and cyclic test are performed on the eBike frame for validation.