The last 50 years saw the development of mechanical simulation thanks to the growth of computers. This development was originally boosted by the industrial production of metal parts. It yielded technologies subsequently applied to new materials such as fiber reinforced plastics with uneven success and consecutive suboptimal designs. Indeed, such materials exhibit much different properties than metals. In particular, their nonlinearity and anisotropy are not straightforward to take into account in simulations. Moreover their properties vary through space (within a given part) and time depending on the manufacturing process and the operating environment. This variability is actually driven by the material microstructure i.e., the amount, shape and orientation adopted by the fibers at the core of the composite. Hence the name of the game in composite material modeling – and the strategy applied by e-Xstream engineering since 10 years – consists in enriching simulations with such microstructural information. This multi-scale modeling strategy will be addressed in this paper, focusing on fiber reinforced plastic durability.
Multi-scale modeling enables accurate structural part life predictions from a limited amount of experimental data. On the one hand, it yields stresses depending on local fiber orientation, required as first input for a structural fatigue analysis. On the other hand, it provides S-N curves for both local microstructure and multiaxial stress state, necessary as second input for the fatigue analysis. Hence, its combination with local fiber orientations from injection molding simulation constitutes a fully coupled approach to lifetime estimation. Indeed, thanks to multi-scale modeling, nCode DesignLife proceeds to stress combination and damage computation based on microstructure-sensitive stresses and S-N curves respectively. The two dependencies on microstructure, embedded in so-called material models, are calibrated on specimen measurements in few directions with respect to the injection direction, routinely performed by Solvay as long as PA66 is concerned. They are subsequently available for structural part life predictions, e.g., for a beam demonstrator employed at Solvay or an engine mount, an automotive part undergoing demanding under-the-hood environmental conditions. Hence they drive reductions in prototyping and design costs related to the structural durability of a component, one of the most expensive attributes to test.
Presenter: Kurt Danielson, e-Xstream engineering
Originally presented on March 5, 2015 at the 2015 HBM-nCode Products User Group Meeting in Livonia, Michigan (USA).