Damage surrogate models for real-time flooding risk assessment of passenger ships

Abstract

Real-time assessment of flooding risk associated with the collision between two ships, requires a fast estimation of damage dimensions and associated survivability. The state-of-the-art frameworks for risk assessment on passenger ships do not consider a direct evaluation of damages through crash simulations but refer to probabilistic considerations, modelling damage characteristics according to statutory marginal distributions of damage breaches too old to be any longer relevant. In any case, such an approach is not possible for the real-time risk assessment process developed in project FLARE, aimed at promoting the employment of first-principles tools for risk evaluation. In this spirit, the present work investigates the possibility of developing a database-oriented damage breach model, which employs direct crash analyses with the super-element code SHARP. However, the sole usage of crash simulations is not suitable for real-time applications. Therefore, starting from the collision simulation database, surrogate models have to be derived for real-time application. In this specific case, three different strategies have been used for the models creation namely: multiple linear regressions, neural networks and decision trees. Here, the strategy to build the database and application on a reference passenger ship is described, highlighting the differences in accuracy and calculation time between the proposed surrogate models.