2.2. Strain-rate effects
Strain-rate effects are not considered herein because the comparisons are made predominantly with quasi-static and simplified test data. However, in a realistic scenario, strain-rate effects may affect the strain localization and thus, in turn, the fracture prediction. Thus, such effects may be important to include in a simulation based on calibrated test data.
Using parameters calibrated to the initial yield stress of, e.g., Cowper-Symonds strain-rate hardening (as in Refs. [5] or [6]) can lead to severe overestimation of the structural stiffness and especially the energy dissipation during a collision, due to the large difference between the effect of strain-rate hardening on initial yield stress as opposed to plastic flow stress; see Törnqvist [7] and Choung et al. [8] for further discussion.
One context in which strain-rate effects are clearly visible is material testing. When diffuse necking is initiated, the velocity of deformation (strain rate) in the neck will increase. Elements that are subjected to elevated strain rates will harden more rapidly than the surrounding elements, thereby stabilizing the development of the neck and extending the zone of strain localizations. The fracturing can be correspondingly delayed. The magnitude of the fracture strain may also be affected by strain-rate effects.
Experiments regarding strain rate exhibit high variability, and it is difficult to obtain a reliable set of parameters to use in a collision simulation without introducing large uncertainty. Furthermore, the effect of reduced fracture strain is even more uncertain; see Jones [9] for a discussion. Care should thus be taken if strain-rate effects are to be included in a full-scale collision simulation.