Probabilistic methods for estimation of the extreme value statistics of ship ice loads

It is well known that when a ship sails in ice-covered regions, the ship-ice interaction process is complex and the associated ice loads on the hull is a stochastic process. Therefore, statistical models and methods should be applied to describe the ice load process. The aim of this work is to present a novel method for estimating the extreme ice loads which is directly related to the reliability of the vessel. This method, briefly referenced to as the ACER (average conditional exceedance rate) method, can provide a reasonable extreme value prediction of the ice loads by efficiently utilizing the available data, which was collected by an ice load monitoring (ILM) system. The basic idea for the ACER approach lies in the fact that a sequence of nonparametric distribution functions are constructed in order to approximate the extreme value distribution of the collected time history. The main principle of the ACER method is presented in detail. Furthermore, the methods based on the classic extreme value theory are also introduced in order to provide a benchmark study.