Abstract

For many years crashworthiness simulation has been – and still is – an extremely demanding field of engineering with an ever-increasing level of complexity when it comes to constitutive or connection modelling. Many new material models for steels, alloys, elastomers and polymers, that are capable of taking various different effects into account (e.g. microstructural evolution due to different chemistry and/or heat treatment), were developed, implemented and successfully applied. Particularly if the local material properties due to pre-straining, reinforcement particles/fibers or heat treatment during part production are dominating the crashworthiness performance, this so-called virtual process chain must be considered in all its complexity as well. While all the aforementioned aspects were regarded and targeted in the past, the spatial discretization often still relies on 5-parameter shell elements based on classical Reissner-Mindlin kinematics. The present contribution will discuss the properties of shells in crashworthiness simulation and present the consequences of these assumptions in the light of material modelling. Clearly, the shown downsides can be compensated by adjusting constitutive parameters – which is the only way to go for models that suffer from too coarse spatial discretization in explicit dynamics. However, another rarely noted consequence is, that in the quest of being as predictive as possible in crashworthiness, i.e. in the postcritical regime of the stress-strain relation in the constitutive model, calibration has to be done for shell and solid discretization independently. The paper will address these topics in detail, exemplify the issues by simple examples and give concise conclusions for everyday work in this interesting field of engineering.