Numerical investigation on a side-shell structure subjected to collision impact load

Abstract

The aim of the present paper is to investigate a correlation between collision damage parameters, which could improve Monte Carlo simulation of random collision damages and facilitate its implementation into the risk assessment of maritime transportation. The proposed approach is based on the explicit LS-Dyna simulation of the collision between a ferry and an Aframax tanker, as striking and struck ship respectively. Numerical simulations are then approximated by the surrogate model enabling evaluation of damage for different random collision scenarios characterized by striking ship speed, collision angle, collision location and displacement mass. Monte Carlo simulation of random collision scenarios is then employed to obtain histograms of damage parameters that can be compared to IMO probabilistic models. Such approach enables not only comparison of damage size histograms between IMO and numerical simulations but also establishing mutual correlation among damage length, breadth and height. Finally, the established correlation is employed into the IMO probabilistic model of random damage parameters and consequences on probabilistic models of residual ultimate strength and still water bending moments are discussed.ship respectively (Parunov et al. 2016). Numerical simulations are then approximated by surrogate model enabling fast computation of damages for different collision scenarios characterized by striking ship speed, collision angle, collision location and displacement mass. MC simulation of random collision scenarios is then employed with surrogate model to obtain histograms of damage parameters that can be compared to IMO probabilistic models. Such approach enables comparison of histograms of damage size between IMO and numerical simulations as well as establishing a mutual correlation between damage length, breadth and height.Correlation is then employed to the IMO probabilistic model of random damage parameters and differences on probability density functions are observed. Consequences of the correlation on the two main pertinent random variables influencing hull girder safety of damaged ship-residual ultimate strength and still water bending moment, are studied. Conclusions about practical importance of correlation among random damage parameters are finally drawn.