Improving container ship panel stress prediction, based on another highly correlated panel stress measurement

Abstract

Operational safety requires that container ships are designed to endure extreme wave loads. This paper studies extreme panel stresses simultaneously measured at two different deck locations of a container vessel sailing in the North Atlantic between Canada and Europe. Ship panel stresses contain whipping part (transient vibratory response of the hull girder due to wave impacts occurring mainly in the bow area), the latter whipping phenomenon is inherently present in on-board stress measurements. It must be noted that measured hydroelastic vessel panel stresses include other complex non-linear effects as well, for example springing.In many practical situations it would be useful to improve accuracy of some statistical predictions extracted from a certain stochastic random process, given another synchronous highly correlated stochastic process that has been measured for a longer time, than the process of interest.The latter situation may happen if one on-board sensor malfunctioned, and it yielded data record much shorter than the neighboring well-functioning sensor. In this paper the latter issue of improving extreme value prediction has been addressed. In other words, this paper studies an efficient transfer of information between two synchronous highly correlated stochastic on-board stress processes. Two correlated ship panel stress processes at mid and aft deck locations were studied in order to test efficiency of the proposed technique of improving one process extreme prediction, based on the other. This paper intends to stimulate further research and fill the gap that is currently visible in the literature on the latter important topic.