On the sensitivity to work hardening and strain-rate effects in nonlinear FEM analysis of ship collisions

Abstract

The complete stress–strain relationship is needed for nonlinear finite element simulation of ship collisions and other events with plastic deformations, whereas only the material grade is known in the design phase. The analyst has to rely on minimum requirements to material strength found in class rules and statistical distributions of the material parameters. This paper investigates the effect of various features of the complete stress–strain curve on the predicted outcome of a collision simulation. This is discussed on the basis of published literature, and the effect of the assumed stress–strain curve is determined through nonlinear finite element analysis (NLFEA) simulations of a full-scale impact scenario. The influence of strain-rate effects is investigated. The findings reveal that the slope of the stress–strain curve determines how strains localise, and thereby when and where fracture propagates. The slope is strongly dependent on the yield ratio, yield plateau and the elongation to fracture, parameters that have significant statistical variation within a material grade. Some challenges related to simulation of strain-rate effects are also discussed.

Keywords:

• crashworthiness
• collision
• work hardening
• strain rate
• NLFEA

Acknowledgements

The numerical investigations were made possible by computational resources at NTNU provided by NOTUR (http://www.notur.no).

The authors would like to acknowledge the many fruitful discussions with Gabriele Notaro, Gunnar Solland and Atle Johansen at DNV-GL that have occurred during the efforts to revise the DNV RP-C208 to include guidance on ship impact simulations. Furthermore, the reviewers are acknowledged for their valuable comments that helped to improve the paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Note that the visco-plastic option for Cowper–Symonds in LS-DYNA is not used; rather the simpler version with a constant DHF is adopted. This allows for comparison with hand-calculations directly. The visco-plastic method is generally recommended (VP=1), and will require re-calibration for dynamic amplification of the plastic flow stress. In general, the strain-rate effect is less for the visco-plastic implementation for the same set of parameters.

2. Note that comparing a rigid stem vs. a deformable side structure has a limited physical relevance as the relative strength is low; hence the stem will be crushed by the side (see Storheim and Amdahl (2014) for further details).

Additional information

Funding

This work was supported in part by the Research Council of Norway through the Centres of Excellence funding scheme, project AMOS [project number 223254]; the Centres for Research-based Innovation scheme, project SAMCoT [project number 203471].